mutter/clutter/clutter-actor.c

15827 lines
467 KiB
C
Raw Normal View History

/*
* Clutter.
*
* An OpenGL based 'interactive canvas' library.
*
* Authored By Matthew Allum <mallum@openedhand.com>
*
* Copyright (C) 2006, 2007, 2008 OpenedHand Ltd
* Copyright (C) 2009, 2010 Intel Corp
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library. If not, see <http://www.gnu.org/licenses/>.
*/
/**
* SECTION:clutter-actor
* @short_description: Base abstract class for all visual stage actors.
*
* The ClutterActor class is the basic element of the scene graph in Clutter,
* and it encapsulates the position, size, and transformations of a node in
* the graph.
*
* <refsect2 id="ClutterActor-transformations">
* <title>Actor transformations</title>
* <para>Each actor can be transformed using methods like
* clutter_actor_set_scale() or clutter_actor_set_rotation(). The order
* in which the transformations are applied is decided by Clutter and it is
* the following:</para>
* <orderedlist>
* <listitem><para>translation by the origin of the #ClutterActor:allocation;</para></listitem>
* <listitem><para>translation by the actor's #ClutterActor:depth;</para></listitem>
* <listitem><para>scaling by the #ClutterActor:scale-x and #ClutterActor:scale-y factors;</para></listitem>
* <listitem><para>rotation around the #ClutterActor:rotation-z-angle and #ClutterActor:rotation-z-center;</para></listitem>
* <listitem><para>rotation around the #ClutterActor:rotation-y-angle and #ClutterActor:rotation-y-center;</para></listitem>
* <listitem><para>rotation around the #ClutterActor:rotation-x-angle and #ClutterActor:rotation-x-center;</para></listitem>
* <listitem><para>negative translation by the #ClutterActor:anchor-x and #ClutterActor:anchor-y point.</para></listitem>
* </orderedlist>
* </refsect2>
*
* <refsect2 id="ClutterActor-geometry">
* <title>Modifying an actor's geometry</title>
* <para>Each actor has a bounding box, called #ClutterActor:allocation
* which is either set by its parent or explicitly through the
* clutter_actor_set_position() and clutter_actor_set_size() methods.
* Each actor also has an implicit preferred size.</para>
* <para>An actors preferred size can be defined by any subclass by
* overriding the #ClutterActorClass.get_preferred_width() and the
* #ClutterActorClass.get_preferred_height() virtual functions, or it can
* be explicitly set by using clutter_actor_set_width() and
* clutter_actor_set_height().</para>
* <para>An actors position can be set explicitly by using
* clutter_actor_set_x() and clutter_actor_set_y(); the coordinates are
* relative to the origin of the actors parent.</para>
* </refsect2>
*
* <refsect2 id="ClutterActor-children">
* <title>Managing actor children</title>
* <para>Each actor can have multiple children, by calling
* clutter_actor_add_child() to add a new child actor, and
* clutter_actor_remove_child() to remove an existing child. #ClutterActor
* will hold a reference on each child actor, which will be released when
* the child is removed from its parent, or destroyed using
* clutter_actor_destroy().</para>
* <informalexample><programlisting>
* ClutterActor *actor = clutter_actor_new ();
*
* /&ast; set the bounding box of the actor &ast;/
* clutter_actor_set_position (actor, 0, 0);
* clutter_actor_set_size (actor, 480, 640);
*
* /&ast; set the background color of the actor &ast;/
* clutter_actor_set_background_color (actor, CLUTTER_COLOR_Orange);
*
* /&ast; set the bounding box of the child, relative to the parent &ast;/
* ClutterActor *child = clutter_actor_new ();
* clutter_actor_set_position (child, 20, 20);
* clutter_actor_set_size (child, 80, 240);
*
* /&ast; set the background color of the child &ast;/
* clutter_actor_set_background_color (child, CLUTTER_COLOR_Blue);
*
* /&ast; add the child to the actor &ast;/
* clutter_actor_add_child (actor, child);
* </programlisting></informalexample>
* <para>Children can be inserted at a given index, or above and below
* another child actor. The order of insertion determines the order of the
* children when iterating over them. Iterating over children is performed
* by using clutter_actor_get_first_child(), clutter_actor_get_previous_sibling(),
2012-01-24 09:36:31 +00:00
* clutter_actor_get_next_sibling(), and clutter_actor_get_last_child(). It is
* also possible to retrieve a list of children by using
* clutter_actor_get_children(), as well as retrieving a specific child at a
* given index by using clutter_actor_get_child_at_index().</para>
* <para>If you need to track additions of children to a #ClutterActor, use
* the #ClutterContainer::actor-added signal; similarly, to track removals
* of children from a ClutterActor, use the #ClutterContainer::actor-removed
* signal.</para>
* <informalexample><programlisting>
* <xi:include xmlns:xi="http://www.w3.org/2001/XInclude" parse="text" href="../../../../tests/interactive/test-actor.c">
* <xi:fallback>FIXME: MISSING XINCLUDE CONTENT</xi:fallback>
* </xi:include>
* </programlisting></informalexample>
* <figure id="bin-layout">
* <title>Actors</title>
* <graphic fileref="test-actor.png" format="PNG"/>
* </figure>
* </refsect2>
*
* <refsect2 id="ClutterActor-events">
* <title>Handling events on an actor</title>
* <para>A #ClutterActor can receive and handle input device events, for
* instance pointer events and key events, as long as its
* #ClutterActor:reactive property is set to %TRUE.</para>
* <para>Once an actor has been determined to be the source of an event,
* Clutter will traverse the scene graph from the top-level actor towards the
* event source, emitting the #ClutterActor::captured-event signal on each
* ancestor until it reaches the source; this phase is also called
* <emphasis>the capture phase</emphasis>. If the event propagation was not
* stopped, the graph is walked backwards, from the source actor to the
* top-level, and the #ClutterActor::event signal, along with other event
* signals if needed, is emitted; this phase is also called <emphasis>the
* bubble phase</emphasis>. At any point of the signal emission, signal
* handlers can stop the propagation through the scene graph by returning
* %CLUTTER_EVENT_STOP; otherwise, they can continue the propagation by
* returning %CLUTTER_EVENT_PROPAGATE.</para>
* </refsect2>
*
* <refsect2 id="ClutterActor-subclassing">
* <title>Implementing an actor</title>
* <para>Careful consideration should be given when deciding to implement
* a #ClutterActor sub-class. It is generally recommended to implement a
* sub-class of #ClutterActor only for actors that should be used as leaf
* nodes of a scene graph.</para>
* <para>If your actor should be painted in a custom way, you should
* override the #ClutterActor::paint signal class handler. You can either
* opt to chain up to the parent class implementation or decide to fully
* override the default paint implementation; Clutter will set up the
* transformations and clip regions prior to emitting the #ClutterActor::paint
* signal.</para>
* <para>By overriding the #ClutterActorClass.get_preferred_width() and
* #ClutterActorClass.get_preferred_height() virtual functions it is
* possible to change or provide the preferred size of an actor; similarly,
* by overriding the #ClutterActorClass.allocate() virtual function it is
* possible to control the layout of the children of an actor. Make sure to
* always chain up to the parent implementation of the
* #ClutterActorClass.allocate() virtual function.</para>
* <para>In general, it is strongly encouraged to use delegation and
* composition instead of direct subclassing.</para>
* </refsect2>
*
* <refsect2 id="ClutterActor-script">
* <title>ClutterActor custom properties for #ClutterScript</title>
* <para>#ClutterActor defines a custom "rotation" property which
* allows a short-hand description of the rotations to be applied
* to an actor.</para>
* <para>The syntax of the "rotation" property is the following:</para>
* <informalexample>
* <programlisting>
* "rotation" : [
* { "&lt;axis&gt;" : [ &lt;angle&gt;, [ &lt;center&gt; ] ] }
* ]
* </programlisting>
* </informalexample>
* <para>where the <emphasis>axis</emphasis> is the name of an enumeration
* value of type #ClutterRotateAxis and <emphasis>angle</emphasis> is a
* floating point value representing the rotation angle on the given axis,
* in degrees.</para>
* <para>The <emphasis>center</emphasis> array is optional, and if present
* it must contain the center of rotation as described by two coordinates:
* Y and Z for "x-axis"; X and Z for "y-axis"; and X and Y for
* "z-axis".</para>
* <para>#ClutterActor will also parse every positional and dimensional
* property defined as a string through clutter_units_from_string(); you
* should read the documentation for the #ClutterUnits parser format for
* the valid units and syntax.</para>
* </refsect2>
*
* <refsect2 id="ClutterActor-animating">
* <title>Custom animatable properties</title>
* <para>#ClutterActor allows accessing properties of #ClutterAction
* and #ClutterConstraint instances associated to an actor instance
* for animation purposes.</para>
* <para>In order to access a specific #ClutterAction or a #ClutterConstraint
* property it is necessary to set the #ClutterActorMeta:name property on the
* given action or constraint.</para>
* <para>The property can be accessed using the following syntax:</para>
* <informalexample>
* <programlisting>
* @&lt;section&gt;.&lt;meta-name&gt;.&lt;property-name&gt;
* </programlisting>
* </informalexample>
* <para>The initial <emphasis>@</emphasis> is mandatory.</para>
* <para>The <emphasis>section</emphasis> fragment can be one between
* "actions", "constraints" and "effects".</para>
* <para>The <emphasis>meta-name</emphasis> fragment is the name of the
* action or constraint, as specified by the #ClutterActorMeta:name
* property.</para>
* <para>The <emphasis>property-name</emphasis> fragment is the name of the
* action or constraint property to be animated.</para>
* <para>The example below animates a #ClutterBindConstraint applied to an
* actor using clutter_actor_animate(). The <emphasis>rect</emphasis> has
* a binding constraint for the <emphasis>origin</emphasis> actor, and in
* its initial state is fully transparent and overlapping the actor to
* which is bound to. </para>
* <informalexample><programlisting>
* constraint = clutter_bind_constraint_new (origin, CLUTTER_BIND_X, 0.0);
* clutter_actor_meta_set_name (CLUTTER_ACTOR_META (constraint), "bind-x");
* clutter_actor_add_constraint (rect, constraint);
*
* constraint = clutter_bind_constraint_new (origin, CLUTTER_BIND_Y, 0.0);
* clutter_actor_meta_set_name (CLUTTER_ACTOR_META (constraint), "bind-y");
* clutter_actor_add_constraint (rect, constraint);
*
* clutter_actor_set_reactive (rect, TRUE);
* clutter_actor_set_opacity (rect, 0);
*
* g_signal_connect (rect, "button-press-event",
* G_CALLBACK (on_button_press),
* NULL);
* </programlisting></informalexample>
* <para>On button press, the rectangle "slides" from behind the actor to
* which is bound to, using the #ClutterBindConstraint:offset property and
* the #ClutterActor:opacity property.</para>
* <informalexample><programlisting>
* float new_offset = clutter_actor_get_width (origin) + h_padding;
*
* clutter_actor_animate (rect, CLUTTER_EASE_OUT_CUBIC, 500,
* "opacity", 255,
* "@constraints.bind-x.offset", new_offset,
* NULL);
* </programlisting></informalexample>
* </refsect2>
*/
/**
* CLUTTER_ACTOR_IS_MAPPED:
* @a: a #ClutterActor
*
* Evaluates to %TRUE if the %CLUTTER_ACTOR_MAPPED flag is set.
*
* The mapped state is set when the actor is visible and all its parents up
* to a top-level (e.g. a #ClutterStage) are visible, realized, and mapped.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* This check can be used to see if an actor is going to be painted, as only
* actors with the %CLUTTER_ACTOR_MAPPED flag set are going to be painted.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* The %CLUTTER_ACTOR_MAPPED flag is managed by Clutter itself, and it should
* not be checked directly; instead, the recommended usage is to connect a
* handler on the #GObject::notify signal for the #ClutterActor:mapped
* property of #ClutterActor, and check the presence of
* the %CLUTTER_ACTOR_MAPPED flag on state changes.
*
* It is also important to note that Clutter may delay the changes of
* the %CLUTTER_ACTOR_MAPPED flag on top-levels due to backend-specific
* limitations, or during the reparenting of an actor, to optimize
* unnecessary (and potentially expensive) state changes.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* Since: 0.2
*/
/**
* CLUTTER_ACTOR_IS_REALIZED:
* @a: a #ClutterActor
*
* Evaluates to %TRUE if the %CLUTTER_ACTOR_REALIZED flag is set.
*
* The realized state has an actor-dependant interpretation. If an
* actor wants to delay allocating resources until it is attached to a
* stage, it may use the realize state to do so. However it is
* perfectly acceptable for an actor to allocate Cogl resources before
* being realized because there is only one drawing context used by Clutter
* so any resources will work on any stage. If an actor is mapped it
* must also be realized, but an actor can be realized and unmapped
* (this is so hiding an actor temporarily doesn't do an expensive
* unrealize/realize).
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* To be realized an actor must be inside a stage, and all its parents
* must be realized.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* Since: 0.2
*/
/**
* CLUTTER_ACTOR_IS_VISIBLE:
* @a: a #ClutterActor
*
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
* Evaluates to %TRUE if the actor has been shown, %FALSE if it's hidden.
* Equivalent to the ClutterActor::visible object property.
*
* Note that an actor is only painted onscreen if it's mapped, which
* means it's visible, and all its parents are visible, and one of the
* parents is a toplevel stage; see also %CLUTTER_ACTOR_IS_MAPPED.
*
* Since: 0.2
*/
/**
* CLUTTER_ACTOR_IS_REACTIVE:
* @a: a #ClutterActor
*
* Evaluates to %TRUE if the %CLUTTER_ACTOR_REACTIVE flag is set.
*
* Only reactive actors will receive event-related signals.
*
* Since: 0.6
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <math.h>
#include "cogl/cogl.h"
#define CLUTTER_DISABLE_DEPRECATION_WARNINGS
#define CLUTTER_ENABLE_EXPERIMENTAL_API
#include "clutter-actor-private.h"
#include "clutter-action.h"
#include "clutter-actor-meta-private.h"
#include "clutter-animatable.h"
#include "clutter-color-static.h"
#include "clutter-color.h"
#include "clutter-constraint.h"
#include "clutter-container.h"
#include "clutter-debug.h"
#include "clutter-effect-private.h"
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
#include "clutter-enum-types.h"
#include "clutter-fixed-layout.h"
#include "clutter-main.h"
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
#include "clutter-marshal.h"
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
#include "clutter-flatten-effect.h"
#include "clutter-paint-volume-private.h"
#include "clutter-private.h"
2009-03-30 15:41:02 +00:00
#include "clutter-profile.h"
#include "clutter-scriptable.h"
#include "clutter-script-private.h"
#include "clutter-stage-private.h"
#include "clutter-units.h"
#include "deprecated/clutter-behaviour.h"
#include "deprecated/clutter-container.h"
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
#define CLUTTER_ACTOR_GET_PRIVATE(obj) \
(G_TYPE_INSTANCE_GET_PRIVATE ((obj), CLUTTER_TYPE_ACTOR, ClutterActorPrivate))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* Internal enum used to control mapped state update. This is a hint
* which indicates when to do something other than just enforce
* invariants.
*/
typedef enum {
MAP_STATE_CHECK, /* just enforce invariants. */
MAP_STATE_MAKE_UNREALIZED, /* force unrealize, ignoring invariants,
* used when about to unparent.
*/
MAP_STATE_MAKE_MAPPED, /* set mapped, error if invariants not met;
* used to set mapped on toplevels.
*/
MAP_STATE_MAKE_UNMAPPED /* set unmapped, even if parent is mapped,
* used just before unmapping parent.
*/
} MapStateChange;
/* 3 entries should be a good compromise, few layout managers
* will ask for 3 different preferred size in each allocation cycle */
#define N_CACHED_SIZE_REQUESTS 3
struct _ClutterActorPrivate
{
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* request mode */
ClutterRequestMode request_mode;
/* our cached size requests for different width / height */
SizeRequest width_requests[N_CACHED_SIZE_REQUESTS];
SizeRequest height_requests[N_CACHED_SIZE_REQUESTS];
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/* An age of 0 means the entry is not set */
guint cached_height_age;
guint cached_width_age;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/* the bounding box of the actor, relative to the parent's
* allocation
*/
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
ClutterActorBox allocation;
ClutterAllocationFlags allocation_flags;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* depth */
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat z;
/* clip, in actor coordinates */
cairo_rectangle_t clip;
/* the cached transformation matrix; see apply_transform() */
CoglMatrix transform;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
guint8 opacity;
gint opacity_override;
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
ClutterOffscreenRedirect offscreen_redirect;
/* This is an internal effect used to implement the
offscreen-redirect property */
ClutterEffect *flatten_effect;
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
/* scene graph */
ClutterActor *parent;
ClutterActor *prev_sibling;
ClutterActor *next_sibling;
ClutterActor *first_child;
ClutterActor *last_child;
gint n_children;
/* tracks whenever the children of an actor are changed; the
* age is incremented by 1 whenever an actor is added or
* removed. the age is not incremented when the first or the
* last child pointers are changed, or when grandchildren of
* an actor are changed.
*/
gint age;
gchar *name; /* a non-unique name, used for debugging */
guint32 id; /* unique id, used for backward compatibility */
gint32 pick_id; /* per-stage unique id, used for picking */
/* a back-pointer to the Pango context that we can use
* to create pre-configured PangoLayout
*/
PangoContext *pango_context;
/* the text direction configured for this child - either by
* application code, or by the actor's parent
*/
ClutterTextDirection text_direction;
/* a counter used to toggle the CLUTTER_INTERNAL_CHILD flag */
gint internal_child;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
/* meta classes */
ClutterMetaGroup *actions;
ClutterMetaGroup *constraints;
ClutterMetaGroup *effects;
/* delegate object used to allocate the children of this actor */
ClutterLayoutManager *layout_manager;
2010-11-18 11:00:08 +00:00
/* used when painting, to update the paint volume */
ClutterEffect *current_effect;
/* This is used to store an effect which needs to be redrawn. A
redraw can be queued to start from a particular effect. This is
used by parametrised effects that can cache an image of the
actor. If a parameter of the effect changes then it only needs to
redraw the cached image, not the actual actor. The pointer is
only valid if is_dirty == TRUE. If the pointer is NULL then the
whole actor is dirty. */
ClutterEffect *effect_to_redraw;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
/* This is used when painting effects to implement the
clutter_actor_continue_paint() function. It points to the node in
the list of effects that is next in the chain */
const GList *next_effect_to_paint;
ClutterPaintVolume paint_volume;
/* NB: This volume isn't relative to this actor, it is in eye
* coordinates so that it can remain valid after the actor changes.
*/
ClutterPaintVolume last_paint_volume;
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
ClutterStageQueueRedrawEntry *queue_redraw_entry;
ClutterColor bg_color;
/* bitfields */
/* fixed position and sizes */
guint position_set : 1;
guint min_width_set : 1;
guint min_height_set : 1;
guint natural_width_set : 1;
guint natural_height_set : 1;
/* cached request is invalid (implies allocation is too) */
guint needs_width_request : 1;
/* cached request is invalid (implies allocation is too) */
guint needs_height_request : 1;
/* cached allocation is invalid (request has changed, probably) */
guint needs_allocation : 1;
guint show_on_set_parent : 1;
guint has_clip : 1;
guint clip_to_allocation : 1;
guint enable_model_view_transform : 1;
guint enable_paint_unmapped : 1;
guint has_pointer : 1;
guint propagated_one_redraw : 1;
guint paint_volume_valid : 1;
guint last_paint_volume_valid : 1;
guint in_clone_paint : 1;
guint transform_valid : 1;
/* This is TRUE if anything has queued a redraw since we were last
painted. In this case effect_to_redraw will point to an effect
the redraw was queued from or it will be NULL if the redraw was
queued without an effect. */
guint is_dirty : 1;
guint bg_color_set : 1;
};
enum
{
PROP_0,
PROP_NAME,
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* X, Y, WIDTH, HEIGHT are "do what I mean" properties;
* when set they force a size request, when gotten they
* get the allocation if the allocation is valid, and the
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* request otherwise
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*/
PROP_X,
PROP_Y,
PROP_WIDTH,
PROP_HEIGHT,
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* Then the rest of these size-related properties are the "actual"
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* underlying properties set or gotten by X, Y, WIDTH, HEIGHT
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*/
PROP_FIXED_X,
PROP_FIXED_Y,
PROP_FIXED_POSITION_SET,
PROP_MIN_WIDTH,
PROP_MIN_WIDTH_SET,
PROP_MIN_HEIGHT,
PROP_MIN_HEIGHT_SET,
PROP_NATURAL_WIDTH,
PROP_NATURAL_WIDTH_SET,
PROP_NATURAL_HEIGHT,
PROP_NATURAL_HEIGHT_SET,
PROP_REQUEST_MODE,
/* Allocation properties are read-only */
PROP_ALLOCATION,
PROP_DEPTH,
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
PROP_CLIP,
PROP_HAS_CLIP,
PROP_CLIP_TO_ALLOCATION,
PROP_OPACITY,
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
PROP_OFFSCREEN_REDIRECT,
PROP_VISIBLE,
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
PROP_MAPPED,
PROP_REALIZED,
PROP_REACTIVE,
PROP_SCALE_X,
PROP_SCALE_Y,
PROP_SCALE_CENTER_X,
PROP_SCALE_CENTER_Y,
PROP_SCALE_GRAVITY,
PROP_ROTATION_ANGLE_X,
PROP_ROTATION_ANGLE_Y,
PROP_ROTATION_ANGLE_Z,
PROP_ROTATION_CENTER_X,
PROP_ROTATION_CENTER_Y,
PROP_ROTATION_CENTER_Z,
/* This property only makes sense for the z rotation because the
others would depend on the actor having a size along the
z-axis */
PROP_ROTATION_CENTER_Z_GRAVITY,
PROP_ANCHOR_X,
PROP_ANCHOR_Y,
PROP_ANCHOR_GRAVITY,
PROP_SHOW_ON_SET_PARENT,
PROP_TEXT_DIRECTION,
PROP_HAS_POINTER,
PROP_ACTIONS,
PROP_CONSTRAINTS,
2010-06-21 09:20:32 +00:00
PROP_EFFECT,
PROP_LAYOUT_MANAGER,
PROP_X_ALIGN,
PROP_Y_ALIGN,
PROP_MARGIN_TOP,
PROP_MARGIN_BOTTOM,
PROP_MARGIN_LEFT,
PROP_MARGIN_RIGHT,
2011-11-20 22:00:01 +00:00
PROP_BACKGROUND_COLOR,
PROP_BACKGROUND_COLOR_SET,
PROP_FIRST_CHILD,
PROP_LAST_CHILD,
2010-06-21 09:20:32 +00:00
PROP_LAST
};
2010-06-21 09:20:32 +00:00
static GParamSpec *obj_props[PROP_LAST];
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
enum
{
SHOW,
HIDE,
DESTROY,
PARENT_SET,
KEY_FOCUS_IN,
KEY_FOCUS_OUT,
PAINT,
PICK,
REALIZE,
UNREALIZE,
QUEUE_REDRAW,
QUEUE_RELAYOUT,
EVENT,
CAPTURED_EVENT,
BUTTON_PRESS_EVENT,
BUTTON_RELEASE_EVENT,
SCROLL_EVENT,
KEY_PRESS_EVENT,
KEY_RELEASE_EVENT,
MOTION_EVENT,
ENTER_EVENT,
LEAVE_EVENT,
ALLOCATION_CHANGED,
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
LAST_SIGNAL
};
static guint actor_signals[LAST_SIGNAL] = { 0, };
static void clutter_container_iface_init (ClutterContainerIface *iface);
static void clutter_scriptable_iface_init (ClutterScriptableIface *iface);
static void clutter_animatable_iface_init (ClutterAnimatableIface *iface);
static void atk_implementor_iface_init (AtkImplementorIface *iface);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* These setters are all static for now, maybe they should be in the
* public API, but they are perhaps obscure enough to leave only as
* properties
*/
static void clutter_actor_set_min_width (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat min_width);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
static void clutter_actor_set_min_height (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat min_height);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
static void clutter_actor_set_natural_width (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat natural_width);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
static void clutter_actor_set_natural_height (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat natural_height);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
static void clutter_actor_set_min_width_set (ClutterActor *self,
gboolean use_min_width);
static void clutter_actor_set_min_height_set (ClutterActor *self,
gboolean use_min_height);
static void clutter_actor_set_natural_width_set (ClutterActor *self,
gboolean use_natural_width);
static void clutter_actor_set_natural_height_set (ClutterActor *self,
gboolean use_natural_height);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
static void clutter_actor_update_map_state (ClutterActor *self,
MapStateChange change);
static void clutter_actor_unrealize_not_hiding (ClutterActor *self);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* Helper routines for managing anchor coords */
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
static void clutter_anchor_coord_get_units (ClutterActor *self,
const AnchorCoord *coord,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat *x,
gfloat *y,
gfloat *z);
static void clutter_anchor_coord_set_units (AnchorCoord *coord,
gfloat x,
gfloat y,
gfloat z);
static ClutterGravity clutter_anchor_coord_get_gravity (const AnchorCoord *coord);
static void clutter_anchor_coord_set_gravity (AnchorCoord *coord,
ClutterGravity gravity);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
static gboolean clutter_anchor_coord_is_zero (const AnchorCoord *coord);
static void _clutter_actor_queue_only_relayout (ClutterActor *self);
static void _clutter_actor_get_relative_transformation_matrix (ClutterActor *self,
ClutterActor *ancestor,
CoglMatrix *matrix);
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
static ClutterPaintVolume *_clutter_actor_get_paint_volume_mutable (ClutterActor *self);
static guint8 clutter_actor_get_paint_opacity_internal (ClutterActor *self);
static void on_layout_manager_changed (ClutterLayoutManager *manager,
ClutterActor *self);
/* Helper macro which translates by the anchor coord, applies the
given transformation and then translates back */
[actor] Allow changing the transformations matrix Currently, the transformation matrix for an actor is constructed from scenegraph-related accessors. An actor, though, can call COGL API to add new transformations inside the paint() implementation, for instance: static void my_foo_paint (ClutterActor *a) { ... cogl_translate (-scroll_x, -scroll_y, 0); ... } Unfortunately these transformations will be completely ignored by the scenegraph machinery; for instance, getting the actor-relative coordinates from event coordinates is going to break badly because of this. In order to make the scenegraph aware of the potential of additional transformations, we need a ::apply_transform() virtual function. This vfunc will pass a CoglMatrix which can be used to apply additional operations: static void my_foo_apply_transform (ClutterActor *a, CoglMatrix *m) { CLUTTER_ACTOR_CLASS (my_foo_parent_class)->apply_transform (a, m); ... cogl_matrix_translate (m, -scroll_x, -scroll_y, 0); ... } The ::paint() implementation will be called with the actor already using the newly applied transformation matrix, as expected: static void my_foo_paint (ClutterActor *a) { ... } The ::apply_transform() implementations *must* chain up, so that the various transformations of each class are preserved. The default implementation inside ClutterActor applies all the transformations defined by the scenegraph-related accessors. Actors performing transformations inside the paint() function will continue to work as previously.
2009-07-01 12:59:13 +00:00
#define TRANSFORM_ABOUT_ANCHOR_COORD(a,m,c,_transform) G_STMT_START { \
gfloat _tx, _ty, _tz; \
clutter_anchor_coord_get_units ((a), (c), &_tx, &_ty, &_tz); \
cogl_matrix_translate ((m), _tx, _ty, _tz); \
{ _transform; } \
cogl_matrix_translate ((m), -_tx, -_ty, -_tz); } G_STMT_END
static GQuark quark_shader_data = 0;
static GQuark quark_actor_layout_info = 0;
static GQuark quark_actor_transform_info = 0;
G_DEFINE_TYPE_WITH_CODE (ClutterActor,
clutter_actor,
G_TYPE_INITIALLY_UNOWNED,
G_IMPLEMENT_INTERFACE (CLUTTER_TYPE_CONTAINER,
clutter_container_iface_init)
G_IMPLEMENT_INTERFACE (CLUTTER_TYPE_SCRIPTABLE,
clutter_scriptable_iface_init)
G_IMPLEMENT_INTERFACE (CLUTTER_TYPE_ANIMATABLE,
clutter_animatable_iface_init)
G_IMPLEMENT_INTERFACE (ATK_TYPE_IMPLEMENTOR,
atk_implementor_iface_init));
/*< private >
* clutter_actor_get_debug_name:
* @actor: a #ClutterActor
*
* Retrieves a printable name of @actor for debugging messages
*
* Return value: a string with a printable name
*/
const gchar *
_clutter_actor_get_debug_name (ClutterActor *actor)
{
return actor->priv->name != NULL ? actor->priv->name
: G_OBJECT_TYPE_NAME (actor);
}
#ifdef CLUTTER_ENABLE_DEBUG
/* XXX - this is for debugging only, remove once working (or leave
* in only in some debug mode). Should leave it for a little while
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
* until we're confident in the new map/realize/visible handling.
*/
static inline void
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
clutter_actor_verify_map_state (ClutterActor *self)
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
{
ClutterActorPrivate *priv = self->priv;
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
if (CLUTTER_ACTOR_IS_REALIZED (self))
{
/* all bets are off during reparent when we're potentially realized,
* but should not be according to invariants
*/
if (!CLUTTER_ACTOR_IN_REPARENT (self))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
if (priv->parent == NULL)
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
if (CLUTTER_ACTOR_IS_TOPLEVEL (self))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
}
else
g_warning ("Realized non-toplevel actor '%s' should "
"have a parent",
_clutter_actor_get_debug_name (self));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
else if (!CLUTTER_ACTOR_IS_REALIZED (priv->parent))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
g_warning ("Realized actor %s has an unrealized parent %s",
_clutter_actor_get_debug_name (self),
_clutter_actor_get_debug_name (priv->parent));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
}
}
if (CLUTTER_ACTOR_IS_MAPPED (self))
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
{
if (!CLUTTER_ACTOR_IS_REALIZED (self))
g_warning ("Actor '%s' is mapped but not realized",
_clutter_actor_get_debug_name (self));
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* remaining bets are off during reparent when we're potentially
* mapped, but should not be according to invariants
*/
if (!CLUTTER_ACTOR_IN_REPARENT (self))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
if (priv->parent == NULL)
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
if (CLUTTER_ACTOR_IS_TOPLEVEL (self))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
if (!CLUTTER_ACTOR_IS_VISIBLE (self) &&
!CLUTTER_ACTOR_IN_DESTRUCTION (self))
{
g_warning ("Toplevel actor '%s' is mapped "
"but not visible",
_clutter_actor_get_debug_name (self));
}
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
else
{
g_warning ("Mapped actor '%s' should have a parent",
_clutter_actor_get_debug_name (self));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
}
else
{
ClutterActor *iter = self;
/* check for the enable_paint_unmapped flag on the actor
* and parents; if the flag is enabled at any point of this
* branch of the scene graph then all the later checks
* become pointless
*/
while (iter != NULL)
{
if (iter->priv->enable_paint_unmapped)
return;
iter = iter->priv->parent;
}
if (!CLUTTER_ACTOR_IS_VISIBLE (priv->parent))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
g_warning ("Actor '%s' should not be mapped if parent '%s'"
"is not visible",
_clutter_actor_get_debug_name (self),
_clutter_actor_get_debug_name (priv->parent));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
if (!CLUTTER_ACTOR_IS_REALIZED (priv->parent))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
g_warning ("Actor '%s' should not be mapped if parent '%s'"
"is not realized",
_clutter_actor_get_debug_name (self),
_clutter_actor_get_debug_name (priv->parent));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
if (!CLUTTER_ACTOR_IS_TOPLEVEL (priv->parent))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
if (!CLUTTER_ACTOR_IS_MAPPED (priv->parent))
g_warning ("Actor '%s' is mapped but its non-toplevel "
"parent '%s' is not mapped",
_clutter_actor_get_debug_name (self),
_clutter_actor_get_debug_name (priv->parent));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
}
}
}
}
#endif /* CLUTTER_ENABLE_DEBUG */
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
static void
clutter_actor_set_mapped (ClutterActor *self,
gboolean mapped)
{
if (CLUTTER_ACTOR_IS_MAPPED (self) == mapped)
return;
if (mapped)
{
CLUTTER_ACTOR_GET_CLASS (self)->map (self);
g_assert (CLUTTER_ACTOR_IS_MAPPED (self));
}
else
{
CLUTTER_ACTOR_GET_CLASS (self)->unmap (self);
g_assert (!CLUTTER_ACTOR_IS_MAPPED (self));
}
}
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* this function updates the mapped and realized states according to
* invariants, in the appropriate order.
*/
static void
clutter_actor_update_map_state (ClutterActor *self,
MapStateChange change)
{
gboolean was_mapped;
was_mapped = CLUTTER_ACTOR_IS_MAPPED (self);
if (CLUTTER_ACTOR_IS_TOPLEVEL (self))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
/* the mapped flag on top-level actors must be set by the
* per-backend implementation because it might be asynchronous.
*
* That is, the MAPPED flag on toplevels currently tracks the X
* server mapped-ness of the window, while the expected behavior
* (if used to GTK) may be to track WM_STATE!=WithdrawnState.
* This creates some weird complexity by breaking the invariant
* that if we're visible and all ancestors shown then we are
* also mapped - instead, we are mapped if all ancestors
* _possibly excepting_ the stage are mapped. The stage
* will map/unmap for example when it is minimized or
* moved to another workspace.
*
* So, the only invariant on the stage is that if visible it
* should be realized, and that it has to be visible to be
* mapped.
*/
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
if (CLUTTER_ACTOR_IS_VISIBLE (self))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
clutter_actor_realize (self);
switch (change)
{
case MAP_STATE_CHECK:
break;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
case MAP_STATE_MAKE_MAPPED:
g_assert (!was_mapped);
clutter_actor_set_mapped (self, TRUE);
break;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
case MAP_STATE_MAKE_UNMAPPED:
g_assert (was_mapped);
clutter_actor_set_mapped (self, FALSE);
break;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
case MAP_STATE_MAKE_UNREALIZED:
/* we only use MAKE_UNREALIZED in unparent,
* and unparenting a stage isn't possible.
* If someone wants to just unrealize a stage
* then clutter_actor_unrealize() doesn't
* go through this codepath.
*/
g_warning ("Trying to force unrealize stage is not allowed");
break;
}
if (CLUTTER_ACTOR_IS_MAPPED (self) &&
!CLUTTER_ACTOR_IS_VISIBLE (self) &&
!CLUTTER_ACTOR_IN_DESTRUCTION (self))
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
{
g_warning ("Clutter toplevel of type '%s' is not visible, but "
"it is somehow still mapped",
_clutter_actor_get_debug_name (self));
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
}
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
else
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
ClutterActorPrivate *priv = self->priv;
ClutterActor *parent = priv->parent;
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
gboolean should_be_mapped;
gboolean may_be_realized;
gboolean must_be_realized;
should_be_mapped = FALSE;
may_be_realized = TRUE;
must_be_realized = FALSE;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (parent == NULL || change == MAP_STATE_MAKE_UNREALIZED)
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
may_be_realized = FALSE;
}
else
{
/* Maintain invariant that if parent is mapped, and we are
* visible, then we are mapped ... unless parent is a
* stage, in which case we map regardless of parent's map
* state but do require stage to be visible and realized.
*
* If parent is realized, that does not force us to be
* realized; but if parent is unrealized, that does force
* us to be unrealized.
*
* The reason we don't force children to realize with
* parents is _clutter_actor_rerealize(); if we require that
* a realized parent means children are realized, then to
* unrealize an actor we would have to unrealize its
* parents, which would end up meaning unrealizing and
* hiding the entire stage. So we allow unrealizing a
* child (as long as that child is not mapped) while that
* child still has a realized parent.
*
* Also, if we unrealize from leaf nodes to root, and
* realize from root to leaf, the invariants are never
* violated if we allow children to be unrealized
* while parents are realized.
*
* When unmapping, MAP_STATE_MAKE_UNMAPPED is specified
* to force us to unmap, even though parent is still
* mapped. This is because we're unmapping from leaf nodes
* up to root nodes.
*/
if (CLUTTER_ACTOR_IS_VISIBLE (self) &&
change != MAP_STATE_MAKE_UNMAPPED)
{
gboolean parent_is_visible_realized_toplevel;
parent_is_visible_realized_toplevel =
(CLUTTER_ACTOR_IS_TOPLEVEL (parent) &&
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
CLUTTER_ACTOR_IS_VISIBLE (parent) &&
CLUTTER_ACTOR_IS_REALIZED (parent));
if (CLUTTER_ACTOR_IS_MAPPED (parent) ||
parent_is_visible_realized_toplevel)
{
must_be_realized = TRUE;
should_be_mapped = TRUE;
}
}
/* if the actor has been set to be painted even if unmapped
* then we should map it and check for realization as well;
* this is an override for the branch of the scene graph
* which begins with this node
*/
if (priv->enable_paint_unmapped)
{
if (priv->parent == NULL)
g_warning ("Attempting to map an unparented actor '%s'",
_clutter_actor_get_debug_name (self));
should_be_mapped = TRUE;
must_be_realized = TRUE;
}
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
if (!CLUTTER_ACTOR_IS_REALIZED (parent))
may_be_realized = FALSE;
}
if (change == MAP_STATE_MAKE_MAPPED && !should_be_mapped)
{
if (parent == NULL)
g_warning ("Attempting to map a child that does not "
"meet the necessary invariants: the actor '%s' "
"has no parent",
_clutter_actor_get_debug_name (self));
else
g_warning ("Attempting to map a child that does not "
"meet the necessary invariants: the actor '%s' "
"is parented to an unmapped actor '%s'",
_clutter_actor_get_debug_name (self),
_clutter_actor_get_debug_name (priv->parent));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
/* If in reparent, we temporarily suspend unmap and unrealize.
*
* We want to go in the order "realize, map" and "unmap, unrealize"
*/
/* Unmap */
if (!should_be_mapped && !CLUTTER_ACTOR_IN_REPARENT (self))
clutter_actor_set_mapped (self, FALSE);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* Realize */
if (must_be_realized)
clutter_actor_realize (self);
/* if we must be realized then we may be, presumably */
g_assert (!(must_be_realized && !may_be_realized));
/* Unrealize */
if (!may_be_realized && !CLUTTER_ACTOR_IN_REPARENT (self))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
clutter_actor_unrealize_not_hiding (self);
/* Map */
if (should_be_mapped)
{
if (!must_be_realized)
g_warning ("Somehow we think actor '%s' should be mapped but "
"not realized, which isn't allowed",
_clutter_actor_get_debug_name (self));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* realization is allowed to fail (though I don't know what
* an app is supposed to do about that - shouldn't it just
* be a g_error? anyway, we have to avoid mapping if this
* happens)
*/
if (CLUTTER_ACTOR_IS_REALIZED (self))
clutter_actor_set_mapped (self, TRUE);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
}
#ifdef CLUTTER_ENABLE_DEBUG
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* check all invariants were kept */
clutter_actor_verify_map_state (self);
#endif
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
static void
clutter_actor_real_map (ClutterActor *self)
{
ClutterActorPrivate *priv = self->priv;
ClutterActor *stage, *iter;
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
g_assert (!CLUTTER_ACTOR_IS_MAPPED (self));
CLUTTER_NOTE (ACTOR, "Mapping actor '%s'",
_clutter_actor_get_debug_name (self));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
CLUTTER_ACTOR_SET_FLAGS (self, CLUTTER_ACTOR_MAPPED);
stage = _clutter_actor_get_stage_internal (self);
priv->pick_id = _clutter_stage_acquire_pick_id (CLUTTER_STAGE (stage), self);
CLUTTER_NOTE (ACTOR, "Pick id '%d' for actor '%s'",
priv->pick_id,
_clutter_actor_get_debug_name (self));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* notify on parent mapped before potentially mapping
* children, so apps see a top-down notification.
*/
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MAPPED]);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
for (iter = self->priv->first_child;
iter != NULL;
iter = iter->priv->next_sibling)
{
clutter_actor_map (iter);
}
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
/**
* clutter_actor_map:
* @self: A #ClutterActor
*
* Sets the %CLUTTER_ACTOR_MAPPED flag on the actor and possibly maps
* and realizes its children if they are visible. Does nothing if the
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
* actor is not visible.
*
* Calling this function is strongly disencouraged: the default
* implementation of #ClutterActorClass.map() will map all the children
* of an actor when mapping its parent.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* When overriding map, it is mandatory to chain up to the parent
* implementation.
*
* Since: 1.0
*/
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
void
clutter_actor_map (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (CLUTTER_ACTOR_IS_MAPPED (self))
return;
if (!CLUTTER_ACTOR_IS_VISIBLE (self))
return;
clutter_actor_update_map_state (self, MAP_STATE_MAKE_MAPPED);
}
static void
clutter_actor_real_unmap (ClutterActor *self)
{
ClutterActorPrivate *priv = self->priv;
ClutterActor *iter;
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
g_assert (CLUTTER_ACTOR_IS_MAPPED (self));
CLUTTER_NOTE (ACTOR, "Unmapping actor '%s'",
_clutter_actor_get_debug_name (self));
for (iter = self->priv->first_child;
iter != NULL;
iter = iter->priv->next_sibling)
{
clutter_actor_unmap (iter);
}
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
CLUTTER_ACTOR_UNSET_FLAGS (self, CLUTTER_ACTOR_MAPPED);
/* clear the contents of the last paint volume, so that hiding + moving +
* showing will not result in the wrong area being repainted
*/
_clutter_paint_volume_init_static (&priv->last_paint_volume, NULL);
priv->last_paint_volume_valid = TRUE;
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* notify on parent mapped after potentially unmapping
* children, so apps see a bottom-up notification.
*/
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MAPPED]);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* relinquish keyboard focus if we were unmapped while owning it */
if (!CLUTTER_ACTOR_IS_TOPLEVEL (self))
{
ClutterStage *stage;
stage = CLUTTER_STAGE (_clutter_actor_get_stage_internal (self));
if (stage != NULL)
_clutter_stage_release_pick_id (stage, priv->pick_id);
priv->pick_id = -1;
if (stage != NULL &&
clutter_stage_get_key_focus (stage) == self)
{
clutter_stage_set_key_focus (stage, NULL);
}
}
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
/**
* clutter_actor_unmap:
* @self: A #ClutterActor
*
* Unsets the %CLUTTER_ACTOR_MAPPED flag on the actor and possibly
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
* unmaps its children if they were mapped.
*
* Calling this function is not encouraged: the default #ClutterActor
* implementation of #ClutterActorClass.unmap() will also unmap any
* eventual children by default when their parent is unmapped.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* When overriding #ClutterActorClass.unmap(), it is mandatory to
* chain up to the parent implementation.
*
* <note>It is important to note that the implementation of the
* #ClutterActorClass.unmap() virtual function may be called after
* the #ClutterActorClass.destroy() or the #GObjectClass.dispose()
* implementation, but it is guaranteed to be called before the
* #GObjectClass.finalize() implementation.</note>
*
* Since: 1.0
*/
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
void
clutter_actor_unmap (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (!CLUTTER_ACTOR_IS_MAPPED (self))
return;
clutter_actor_update_map_state (self, MAP_STATE_MAKE_UNMAPPED);
}
static void
clutter_actor_real_show (ClutterActor *self)
{
if (!CLUTTER_ACTOR_IS_VISIBLE (self))
{
ClutterActorPrivate *priv = self->priv;
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
CLUTTER_ACTOR_SET_FLAGS (self, CLUTTER_ACTOR_VISIBLE);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* we notify on the "visible" flag in the clutter_actor_show()
* wrapper so the entire show signal emission completes first
* (?)
*/
clutter_actor_update_map_state (self, MAP_STATE_CHECK);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* we queue a relayout unless the actor is inside a
* container that explicitly told us not to
*/
if (priv->parent != NULL &&
(!(priv->parent->flags & CLUTTER_ACTOR_NO_LAYOUT)))
{
/* While an actor is hidden the parent may not have
* allocated/requested so we need to start from scratch
* and avoid the short-circuiting in
* clutter_actor_queue_relayout().
*/
priv->needs_width_request = FALSE;
priv->needs_height_request = FALSE;
priv->needs_allocation = FALSE;
clutter_actor_queue_relayout (self);
}
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
}
}
static inline void
set_show_on_set_parent (ClutterActor *self,
gboolean set_show)
{
ClutterActorPrivate *priv = self->priv;
set_show = !!set_show;
if (priv->show_on_set_parent == set_show)
return;
if (priv->parent == NULL)
{
priv->show_on_set_parent = set_show;
g_object_notify_by_pspec (G_OBJECT (self),
obj_props[PROP_SHOW_ON_SET_PARENT]);
}
}
/**
* clutter_actor_show:
* @self: A #ClutterActor
*
* Flags an actor to be displayed. An actor that isn't shown will not
* be rendered on the stage.
*
* Actors are visible by default.
*
* If this function is called on an actor without a parent, the
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* #ClutterActor:show-on-set-parent will be set to %TRUE as a side
* effect.
*/
void
clutter_actor_show (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
/* simple optimization */
if (CLUTTER_ACTOR_IS_VISIBLE (self))
{
/* we still need to set the :show-on-set-parent property, in
* case show() is called on an unparented actor
*/
set_show_on_set_parent (self, TRUE);
return;
}
#ifdef CLUTTER_ENABLE_DEBUG
clutter_actor_verify_map_state (self);
#endif
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
priv = self->priv;
g_object_freeze_notify (G_OBJECT (self));
set_show_on_set_parent (self, TRUE);
g_signal_emit (self, actor_signals[SHOW], 0);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_VISIBLE]);
if (priv->parent != NULL)
clutter_actor_queue_redraw (priv->parent);
g_object_thaw_notify (G_OBJECT (self));
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
}
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
/**
* clutter_actor_show_all:
* @self: a #ClutterActor
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Calls clutter_actor_show() on all children of an actor (if any).
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
*
* Since: 0.2
*
* Deprecated: 1.10: Actors are visible by default
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
*/
void
clutter_actor_show_all (ClutterActor *self)
{
ClutterActorClass *klass;
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
klass = CLUTTER_ACTOR_GET_CLASS (self);
if (klass->show_all)
klass->show_all (self);
}
static void
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
clutter_actor_real_hide (ClutterActor *self)
{
if (CLUTTER_ACTOR_IS_VISIBLE (self))
{
ClutterActorPrivate *priv = self->priv;
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
CLUTTER_ACTOR_UNSET_FLAGS (self, CLUTTER_ACTOR_VISIBLE);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* we notify on the "visible" flag in the clutter_actor_hide()
* wrapper so the entire hide signal emission completes first
* (?)
*/
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
clutter_actor_update_map_state (self, MAP_STATE_CHECK);
/* we queue a relayout unless the actor is inside a
* container that explicitly told us not to
*/
if (priv->parent != NULL &&
(!(priv->parent->flags & CLUTTER_ACTOR_NO_LAYOUT)))
clutter_actor_queue_relayout (priv->parent);
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
}
}
/**
* clutter_actor_hide:
* @self: A #ClutterActor
*
* Flags an actor to be hidden. A hidden actor will not be
* rendered on the stage.
*
* Actors are visible by default.
*
* If this function is called on an actor without a parent, the
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* #ClutterActor:show-on-set-parent property will be set to %FALSE
* as a side-effect.
*/
void
clutter_actor_hide (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
/* simple optimization */
if (!CLUTTER_ACTOR_IS_VISIBLE (self))
{
/* we still need to set the :show-on-set-parent property, in
* case hide() is called on an unparented actor
*/
set_show_on_set_parent (self, FALSE);
return;
}
#ifdef CLUTTER_ENABLE_DEBUG
clutter_actor_verify_map_state (self);
#endif
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
priv = self->priv;
g_object_freeze_notify (G_OBJECT (self));
set_show_on_set_parent (self, FALSE);
g_signal_emit (self, actor_signals[HIDE], 0);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_VISIBLE]);
if (priv->parent != NULL)
clutter_actor_queue_redraw (priv->parent);
g_object_thaw_notify (G_OBJECT (self));
}
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
/**
* clutter_actor_hide_all:
* @self: a #ClutterActor
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Calls clutter_actor_hide() on all child actors (if any).
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
*
* Since: 0.2
*
* Deprecated: 1.10: Using clutter_actor_hide() on the actor will
* prevent its children from being painted as well.
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
*/
void
clutter_actor_hide_all (ClutterActor *self)
{
ClutterActorClass *klass;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
klass = CLUTTER_ACTOR_GET_CLASS (self);
if (klass->hide_all)
klass->hide_all (self);
}
/**
* clutter_actor_realize:
* @self: A #ClutterActor
*
* Realization informs the actor that it is attached to a stage. It
* can use this to allocate resources if it wanted to delay allocation
* until it would be rendered. However it is perfectly acceptable for
* an actor to create resources before being realized because Clutter
* only ever has a single rendering context so that actor is free to
* be moved from one stage to another.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* This function does nothing if the actor is already realized.
*
* Because a realized actor must have realized parent actors, calling
* clutter_actor_realize() will also realize all parents of the actor.
*
* This function does not realize child actors, except in the special
* case that realizing the stage, when the stage is visible, will
* suddenly map (and thus realize) the children of the stage.
**/
void
clutter_actor_realize (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
#ifdef CLUTTER_ENABLE_DEBUG
clutter_actor_verify_map_state (self);
#endif
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
if (CLUTTER_ACTOR_IS_REALIZED (self))
return;
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* To be realized, our parent actors must be realized first.
* This will only succeed if we're inside a toplevel.
*/
if (priv->parent != NULL)
clutter_actor_realize (priv->parent);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
if (CLUTTER_ACTOR_IS_TOPLEVEL (self))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
{
/* toplevels can be realized at any time */
}
else
{
/* "Fail" the realization if parent is missing or unrealized;
* this should really be a g_warning() not some kind of runtime
* failure; how can an app possibly recover? Instead it's a bug
* in the app and the app should get an explanatory warning so
* someone can fix it. But for now it's too hard to fix this
* because e.g. ClutterTexture needs reworking.
*/
if (priv->parent == NULL ||
!CLUTTER_ACTOR_IS_REALIZED (priv->parent))
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
return;
}
CLUTTER_NOTE (ACTOR, "Realizing actor '%s'", _clutter_actor_get_debug_name (self));
2009-05-11 11:36:14 +00:00
CLUTTER_ACTOR_SET_FLAGS (self, CLUTTER_ACTOR_REALIZED);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_REALIZED]);
g_signal_emit (self, actor_signals[REALIZE], 0);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* Stage actor is allowed to unset the realized flag again in its
* default signal handler, though that is a pathological situation.
*/
/* If realization "failed" we'll have to update child state. */
clutter_actor_update_map_state (self, MAP_STATE_CHECK);
}
static void
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
clutter_actor_real_unrealize (ClutterActor *self)
{
/* we must be unmapped (implying our children are also unmapped) */
g_assert (!CLUTTER_ACTOR_IS_MAPPED (self));
}
/**
* clutter_actor_unrealize:
* @self: A #ClutterActor
*
* Unrealization informs the actor that it may be being destroyed or
* moved to another stage. The actor may want to destroy any
* underlying graphics resources at this point. However it is
* perfectly acceptable for it to retain the resources until the actor
* is destroyed because Clutter only ever uses a single rendering
* context and all of the graphics resources are valid on any stage.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* Because mapped actors must be realized, actors may not be
* unrealized if they are mapped. This function hides the actor to be
* sure it isn't mapped, an application-visible side effect that you
* may not be expecting.
*
2010-08-17 13:54:20 +00:00
* This function should not be called by application code.
*/
void
clutter_actor_unrealize (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (!CLUTTER_ACTOR_IS_MAPPED (self));
/* This function should not really be in the public API, because
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
* there isn't a good reason to call it. ClutterActor will already
* unrealize things for you when it's important to do so.
*
* If you were using clutter_actor_unrealize() in a dispose
* implementation, then don't, just chain up to ClutterActor's
* dispose.
*
* If you were using clutter_actor_unrealize() to implement
* unrealizing children of your container, then don't, ClutterActor
* will already take care of that.
*
* If you were using clutter_actor_unrealize() to re-realize to
* create your resources in a different way, then use
* _clutter_actor_rerealize() (inside Clutter) or just call your
* code that recreates your resources directly (outside Clutter).
*/
#ifdef CLUTTER_ENABLE_DEBUG
clutter_actor_verify_map_state (self);
#endif
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
clutter_actor_hide (self);
clutter_actor_unrealize_not_hiding (self);
}
static ClutterActorTraverseVisitFlags
unrealize_actor_before_children_cb (ClutterActor *self,
int depth,
void *user_data)
{
/* If an actor is already unrealized we know its children have also
* already been unrealized... */
if (!CLUTTER_ACTOR_IS_REALIZED (self))
return CLUTTER_ACTOR_TRAVERSE_VISIT_SKIP_CHILDREN;
g_signal_emit (self, actor_signals[UNREALIZE], 0);
return CLUTTER_ACTOR_TRAVERSE_VISIT_CONTINUE;
}
static ClutterActorTraverseVisitFlags
unrealize_actor_after_children_cb (ClutterActor *self,
int depth,
void *user_data)
{
/* We want to unset the realized flag only _after_
* child actors are unrealized, to maintain invariants.
*/
CLUTTER_ACTOR_UNSET_FLAGS (self, CLUTTER_ACTOR_REALIZED);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_REALIZED]);
return CLUTTER_ACTOR_TRAVERSE_VISIT_CONTINUE;
}
/*
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
* clutter_actor_unrealize_not_hiding:
* @self: A #ClutterActor
*
* Unrealization informs the actor that it may be being destroyed or
* moved to another stage. The actor may want to destroy any
* underlying graphics resources at this point. However it is
* perfectly acceptable for it to retain the resources until the actor
* is destroyed because Clutter only ever uses a single rendering
* context and all of the graphics resources are valid on any stage.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* Because mapped actors must be realized, actors may not be
* unrealized if they are mapped. You must hide the actor or one of
* its parents before attempting to unrealize.
*
* This function is separate from clutter_actor_unrealize() because it
* does not automatically hide the actor.
* Actors need not be hidden to be unrealized, they just need to
* be unmapped. In fact we don't want to mess up the application's
* setting of the "visible" flag, so hiding is very undesirable.
*
* clutter_actor_unrealize() does a clutter_actor_hide() just for
* backward compatibility.
*/
static void
clutter_actor_unrealize_not_hiding (ClutterActor *self)
{
_clutter_actor_traverse (self,
CLUTTER_ACTOR_TRAVERSE_DEPTH_FIRST,
unrealize_actor_before_children_cb,
unrealize_actor_after_children_cb,
NULL);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
}
/*
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
* _clutter_actor_rerealize:
* @self: A #ClutterActor
* @callback: Function to call while unrealized
* @data: data for callback
*
* If an actor is already unrealized, this just calls the callback.
*
* If it is realized, it unrealizes temporarily, calls the callback,
* and then re-realizes the actor.
*
* As a side effect, leaves all children of the actor unrealized if
* the actor was realized but not showing. This is because when we
* unrealize the actor temporarily we must unrealize its children
* (e.g. children of a stage can't be realized if stage window is
* gone). And we aren't clever enough to save the realization state of
* all children. In most cases this should not matter, because
* the children will automatically realize when they next become mapped.
*/
void
_clutter_actor_rerealize (ClutterActor *self,
ClutterCallback callback,
void *data)
{
gboolean was_mapped;
gboolean was_showing;
gboolean was_realized;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
#ifdef CLUTTER_ENABLE_DEBUG
clutter_actor_verify_map_state (self);
#endif
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
was_realized = CLUTTER_ACTOR_IS_REALIZED (self);
was_mapped = CLUTTER_ACTOR_IS_MAPPED (self);
was_showing = CLUTTER_ACTOR_IS_VISIBLE (self);
/* Must be unmapped to unrealize. Note we only have to hide this
* actor if it was mapped (if all parents were showing). If actor
* is merely visible (but not mapped), then that's fine, we can
* leave it visible.
*/
if (was_mapped)
clutter_actor_hide (self);
g_assert (!CLUTTER_ACTOR_IS_MAPPED (self));
/* unrealize self and all children */
clutter_actor_unrealize_not_hiding (self);
if (callback != NULL)
{
(* callback) (self, data);
}
if (was_showing)
clutter_actor_show (self); /* will realize only if mapping implies it */
else if (was_realized)
clutter_actor_realize (self); /* realize self and all parents */
}
static void
clutter_actor_real_pick (ClutterActor *self,
const ClutterColor *color)
{
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* the default implementation is just to paint a rectangle
* with the same size of the actor using the passed color
*/
if (clutter_actor_should_pick_paint (self))
{
ClutterActorBox box = { 0, };
float width, height;
clutter_actor_get_allocation_box (self, &box);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
width = box.x2 - box.x1;
height = box.y2 - box.y1;
2008-11-12 Emmanuele Bassi <ebassi@linux.intel.com> * clutter/cogl/cogl-color.h: * clutter/cogl/cogl-path.h: * clutter/cogl/cogl-types.h: * clutter/cogl/common/cogl-color.c: Deprecated cogl_color() in favour of cogl_set_source_color() and friends; store the CoglColor components as unsigned bytes instead of fixed point normalized values; add functions for allocating, copying and freeing CoglColor, for use of language bindings. * clutter/cogl/cogl.h.in: * clutter/cogl/cogl-deprecated.h: Added cogl-deprecated.h, an header file containing the deprecation symbols similar to clutter-deprecated.h. * clutter/cogl/gl/Makefile.am: * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/Makefile.am: * clutter/cogl/gles/cogl-texture.c: * clutter/cogl/gles/cogl.c: Update the GL and GLES implementations of COGL after the CoglColor changes. * clutter/clutter-actor.c: * clutter/clutter-clone-texture.c: * clutter/clutter-entry.c: * clutter/clutter-label.c: * clutter/clutter-rectangle.c: * clutter/clutter-texture.c: Do not use CoglColor whenever it is possible, and use cogl_set_source_color4ub() instead. * clutter/pango/cogl-pango-render.c: Ditto as above. * doc/reference/clutter/subclassing-ClutterActor.xml: * doc/reference/cogl/cogl-sections.txt: Update the documentation. * tests/interactive/test-cogl-offscreen.c: * tests/interactive/test-cogl-primitives.c: * tests/interactive/test-cogl-tex-convert.c: * tests/interactive/test-cogl-tex-foreign.c: * tests/interactive/test-cogl-tex-getset.c: * tests/interactive/test-cogl-tex-polygon.c: * tests/interactive/test-cogl-tex-tile.c: * tests/interactive/test-paint-wrapper.c: Drop the usage of CoglColor whenever it is possible.
2008-11-12 13:57:58 +00:00
cogl_set_source_color4ub (color->red,
color->green,
color->blue,
color->alpha);
cogl_rectangle (0, 0, width, height);
}
/* XXX - this thoroughly sucks, but we need to maintain compatibility
* with existing container classes that override the pick() virtual
* and chain up to the default implementation - otherwise we'll end up
* painting our children twice.
*
* this has to go away for 2.0; hopefully along the pick() itself.
*/
if (CLUTTER_ACTOR_GET_CLASS (self)->pick == clutter_actor_real_pick)
{
ClutterActor *iter;
for (iter = self->priv->first_child;
iter != NULL;
iter = iter->priv->next_sibling)
clutter_actor_paint (iter);
}
}
/**
* clutter_actor_should_pick_paint:
* @self: A #ClutterActor
*
* Should be called inside the implementation of the
* #ClutterActor::pick virtual function in order to check whether
* the actor should paint itself in pick mode or not.
*
* This function should never be called directly by applications.
*
* Return value: %TRUE if the actor should paint its silhouette,
* %FALSE otherwise
*/
gboolean
clutter_actor_should_pick_paint (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
if (CLUTTER_ACTOR_IS_MAPPED (self) &&
(_clutter_context_get_pick_mode () == CLUTTER_PICK_ALL ||
CLUTTER_ACTOR_IS_REACTIVE (self)))
return TRUE;
return FALSE;
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
static void
clutter_actor_real_get_preferred_width (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat for_height,
gfloat *min_width_p,
gfloat *natural_width_p)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
ClutterActorPrivate *priv = self->priv;
if (priv->n_children != 0 &&
priv->layout_manager != NULL)
{
ClutterContainer *container = CLUTTER_CONTAINER (self);
CLUTTER_NOTE (LAYOUT, "Querying the layout manager '%s'[%p] "
"for the preferred width",
G_OBJECT_TYPE_NAME (priv->layout_manager),
priv->layout_manager);
clutter_layout_manager_get_preferred_width (priv->layout_manager,
container,
for_height,
min_width_p,
natural_width_p);
return;
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* Default implementation is always 0x0, usually an actor
* using this default is relying on someone to set the
* request manually
*/
CLUTTER_NOTE (LAYOUT, "Default preferred width: 0, 0");
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (min_width_p)
*min_width_p = 0;
if (natural_width_p)
*natural_width_p = 0;
}
static void
clutter_actor_real_get_preferred_height (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat for_width,
gfloat *min_height_p,
gfloat *natural_height_p)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
ClutterActorPrivate *priv = self->priv;
if (priv->n_children != 0 &&
priv->layout_manager != NULL)
{
ClutterContainer *container = CLUTTER_CONTAINER (self);
CLUTTER_NOTE (LAYOUT, "Querying the layout manager '%s'[%p] "
"for the preferred height",
G_OBJECT_TYPE_NAME (priv->layout_manager),
priv->layout_manager);
clutter_layout_manager_get_preferred_height (priv->layout_manager,
container,
for_width,
min_height_p,
natural_height_p);
return;
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* Default implementation is always 0x0, usually an actor
* using this default is relying on someone to set the
* request manually
*/
CLUTTER_NOTE (LAYOUT, "Default preferred height: 0, 0");
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (min_height_p)
*min_height_p = 0;
if (natural_height_p)
*natural_height_p = 0;
}
static void
clutter_actor_store_old_geometry (ClutterActor *self,
ClutterActorBox *box)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
*box = self->priv->allocation;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
}
static inline void
clutter_actor_notify_if_geometry_changed (ClutterActor *self,
const ClutterActorBox *old)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
ClutterActorPrivate *priv = self->priv;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
GObject *obj = G_OBJECT (self);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_object_freeze_notify (obj);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* to avoid excessive requisition or allocation cycles we
* use the cached values.
*
* - if we don't have an allocation we assume that we need
* to notify anyway
* - if we don't have a width or a height request we notify
* width and height
* - if we have a valid allocation then we check the old
* bounding box with the current allocation and we notify
* the changes
*/
if (priv->needs_allocation)
{
g_object_notify_by_pspec (obj, obj_props[PROP_X]);
g_object_notify_by_pspec (obj, obj_props[PROP_Y]);
g_object_notify_by_pspec (obj, obj_props[PROP_WIDTH]);
g_object_notify_by_pspec (obj, obj_props[PROP_HEIGHT]);
}
else if (priv->needs_width_request || priv->needs_height_request)
{
g_object_notify_by_pspec (obj, obj_props[PROP_WIDTH]);
g_object_notify_by_pspec (obj, obj_props[PROP_HEIGHT]);
}
else
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat xu, yu;
gfloat widthu, heightu;
xu = priv->allocation.x1;
yu = priv->allocation.y1;
widthu = priv->allocation.x2 - priv->allocation.x1;
heightu = priv->allocation.y2 - priv->allocation.y1;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (xu != old->x1)
g_object_notify_by_pspec (obj, obj_props[PROP_X]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (yu != old->y1)
g_object_notify_by_pspec (obj, obj_props[PROP_Y]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (widthu != (old->x2 - old->x1))
g_object_notify_by_pspec (obj, obj_props[PROP_WIDTH]);
if (heightu != (old->y2 - old->y1))
g_object_notify_by_pspec (obj, obj_props[PROP_HEIGHT]);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_object_thaw_notify (obj);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
}
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
/*< private >
* clutter_actor_set_allocation_internal:
* @self: a #ClutterActor
* @box: a #ClutterActorBox
* @flags: allocation flags
*
* Stores the allocation of @self.
*
* This function only performs basic storage and property notification.
*
* This function should be called by clutter_actor_set_allocation()
* and by the default implementation of #ClutterActorClass.allocate().
*
* Return value: %TRUE if the allocation of the #ClutterActor has been
* changed, and %FALSE otherwise
*/
static inline gboolean
clutter_actor_set_allocation_internal (ClutterActor *self,
const ClutterActorBox *box,
ClutterAllocationFlags flags)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
ClutterActorPrivate *priv = self->priv;
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
GObject *obj;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
gboolean x1_changed, y1_changed, x2_changed, y2_changed;
gboolean flags_changed;
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
gboolean retval;
ClutterActorBox old_alloc = { 0, };
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
obj = G_OBJECT (self);
g_object_freeze_notify (obj);
clutter_actor_store_old_geometry (self, &old_alloc);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
x1_changed = priv->allocation.x1 != box->x1;
y1_changed = priv->allocation.y1 != box->y1;
x2_changed = priv->allocation.x2 != box->x2;
y2_changed = priv->allocation.y2 != box->y2;
flags_changed = priv->allocation_flags != flags;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->allocation = *box;
priv->allocation_flags = flags;
/* allocation is authoritative */
priv->needs_width_request = FALSE;
priv->needs_height_request = FALSE;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->needs_allocation = FALSE;
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
if (x1_changed || y1_changed || x2_changed || y2_changed || flags_changed)
{
CLUTTER_NOTE (LAYOUT, "Allocation for '%s' changed",
_clutter_actor_get_debug_name (self));
priv->transform_valid = FALSE;
g_object_notify_by_pspec (obj, obj_props[PROP_ALLOCATION]);
retval = TRUE;
}
else
retval = FALSE;
clutter_actor_notify_if_geometry_changed (self, &old_alloc);
g_object_thaw_notify (obj);
return retval;
}
static void clutter_actor_real_allocate (ClutterActor *self,
const ClutterActorBox *box,
ClutterAllocationFlags flags);
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
static inline void
clutter_actor_maybe_layout_children (ClutterActor *self,
const ClutterActorBox *allocation,
ClutterAllocationFlags flags)
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
{
ClutterActorPrivate *priv = self->priv;
/* this is going to be a bit hard to follow, so let's put an explanation
* here.
*
* we want ClutterActor to have a default layout manager if the actor was
* created using "g_object_new (CLUTTER_TYPE_ACTOR, NULL)".
*
* we also want any subclass of ClutterActor that does not override the
* ::allocate() virtual function to delegate to a layout manager.
*
* finally, we want to allow people subclassing ClutterActor and overriding
* the ::allocate() vfunc to let Clutter delegate to the layout manager.
*
* on the other hand, we want existing actor subclasses overriding the
* ::allocate() virtual function and chaining up to the parent's
* implementation to continue working without allocating their children
* twice, or without entering an allocation loop.
*
* for the first two points, we check if the class of the actor is
* overridding the ::allocate() virtual function; if it isn't, then we
* follow through with checking whether we have children and a layout
* manager, and eventually calling clutter_layout_manager_allocate().
*
* for the third point, we check the CLUTTER_DELEGATE_LAYOUT flag in the
* allocation flags that we got passed, and if it is present, we continue
* with the check above.
*
* if neither of these two checks yields a positive result, we just
* assume that the ::allocate() virtual function that resulted in this
* function being called will also allocate the children of the actor.
*/
if (CLUTTER_ACTOR_GET_CLASS (self)->allocate == clutter_actor_real_allocate)
goto check_layout;
if ((flags & CLUTTER_DELEGATE_LAYOUT) != 0)
goto check_layout;
return;
check_layout:
if (priv->n_children != 0 &&
priv->layout_manager != NULL)
{
ClutterContainer *container = CLUTTER_CONTAINER (self);
ClutterAllocationFlags children_flags;
ClutterActorBox children_box;
/* normalize the box passed to the layout manager */
children_box.x1 = children_box.y1 = 0.f;
children_box.x2 = (allocation->x2 - allocation->x1);
children_box.y2 = (allocation->y2 - allocation->y1);
/* remove the DELEGATE_LAYOUT flag; this won't be passed to
* the actor's children, since it refers only to the current
* actor's allocation.
*/
children_flags = flags;
children_flags &= ~CLUTTER_DELEGATE_LAYOUT;
2011-12-20 14:47:04 +00:00
CLUTTER_NOTE (LAYOUT,
"Allocating %d children of %s "
"at { %.2f, %.2f - %.2f x %.2f } "
"using %s",
priv->n_children,
2011-12-20 14:47:04 +00:00
_clutter_actor_get_debug_name (self),
allocation->x1,
allocation->y1,
(allocation->x2 - allocation->x1),
(allocation->y2 - allocation->y1),
2011-12-20 14:47:04 +00:00
G_OBJECT_TYPE_NAME (priv->layout_manager));
clutter_layout_manager_allocate (priv->layout_manager,
container,
&children_box,
children_flags);
}
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
}
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
static void
clutter_actor_real_allocate (ClutterActor *self,
const ClutterActorBox *box,
ClutterAllocationFlags flags)
{
ClutterActorPrivate *priv = self->priv;
gboolean changed;
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
g_object_freeze_notify (G_OBJECT (self));
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
changed = clutter_actor_set_allocation_internal (self, box, flags);
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
/* we allocate our children before we notify changes in our geometry,
* so that people connecting to properties will be able to get valid
* data out of the sub-tree of the scene graph that has this actor at
* the root.
*/
clutter_actor_maybe_layout_children (self, box, flags);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
if (changed)
g_signal_emit (self, actor_signals[ALLOCATION_CHANGED], 0,
&priv->allocation,
priv->allocation_flags);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_thaw_notify (G_OBJECT (self));
}
static void
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
_clutter_actor_signal_queue_redraw (ClutterActor *self,
ClutterActor *origin)
{
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
/* no point in queuing a redraw on a destroyed actor */
if (CLUTTER_ACTOR_IN_DESTRUCTION (self))
return;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
/* NB: We can't bail out early here if the actor is hidden in case
* the actor bas been cloned. In this case the clone will need to
* receive the signal so it can queue its own redraw.
*/
/* calls klass->queue_redraw in default handler */
g_signal_emit (self, actor_signals[QUEUE_REDRAW], 0, origin);
}
static void
clutter_actor_real_queue_redraw (ClutterActor *self,
ClutterActor *origin)
{
ClutterActor *parent;
CLUTTER_NOTE (PAINT, "Redraw queued on '%s' (from: '%s')",
_clutter_actor_get_debug_name (self),
origin != NULL ? _clutter_actor_get_debug_name (origin)
: "same actor");
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
/* no point in queuing a redraw on a destroyed actor */
if (CLUTTER_ACTOR_IN_DESTRUCTION (self))
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
return;
/* If the queue redraw is coming from a child then the actor has
become dirty and any queued effect is no longer valid */
if (self != origin)
{
self->priv->is_dirty = TRUE;
self->priv->effect_to_redraw = NULL;
}
/* If the actor isn't visible, we still had to emit the signal
* to allow for a ClutterClone, but the appearance of the parent
* won't change so we don't have to propagate up the hierarchy.
*/
if (!CLUTTER_ACTOR_IS_VISIBLE (self))
return;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
/* Although we could determine here that a full stage redraw
* has already been queued and immediately bail out, we actually
* guarantee that we will propagate a queue-redraw signal to our
* parent at least once so that it's possible to implement a
* container that tracks which of its children have queued a
* redraw.
*/
if (self->priv->propagated_one_redraw)
{
ClutterActor *stage = _clutter_actor_get_stage_internal (self);
if (stage != NULL &&
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
_clutter_stage_has_full_redraw_queued (CLUTTER_STAGE (stage)))
return;
}
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
self->priv->propagated_one_redraw = TRUE;
/* notify parents, if they are all visible eventually we'll
* queue redraw on the stage, which queues the redraw idle.
*/
parent = clutter_actor_get_parent (self);
if (parent != NULL)
{
/* this will go up recursively */
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
_clutter_actor_signal_queue_redraw (parent, origin);
}
}
static void
clutter_actor_real_queue_relayout (ClutterActor *self)
{
ClutterActorPrivate *priv = self->priv;
/* no point in queueing a redraw on a destroyed actor */
if (CLUTTER_ACTOR_IN_DESTRUCTION (self))
return;
priv->needs_width_request = TRUE;
priv->needs_height_request = TRUE;
priv->needs_allocation = TRUE;
/* reset the cached size requests */
memset (priv->width_requests, 0,
N_CACHED_SIZE_REQUESTS * sizeof (SizeRequest));
memset (priv->height_requests, 0,
N_CACHED_SIZE_REQUESTS * sizeof (SizeRequest));
/* We need to go all the way up the hierarchy */
if (priv->parent != NULL)
_clutter_actor_queue_only_relayout (priv->parent);
}
/**
* clutter_actor_apply_relative_transform_to_point:
* @self: A #ClutterActor
2009-07-02 13:21:16 +00:00
* @ancestor: (allow-none): A #ClutterActor ancestor, or %NULL to use the
* default #ClutterStage
* @point: A point as #ClutterVertex
* @vertex: (out caller-allocates): The translated #ClutterVertex
*
* Transforms @point in coordinates relative to the actor into
* ancestor-relative coordinates using the relevant transform
* stack (i.e. scale, rotation, etc).
*
* If @ancestor is %NULL the ancestor will be the #ClutterStage. In
* this case, the coordinates returned will be the coordinates on
* the stage before the projection is applied. This is different from
* the behaviour of clutter_actor_apply_transform_to_point().
*
* Since: 0.6
*/
void
clutter_actor_apply_relative_transform_to_point (ClutterActor *self,
ClutterActor *ancestor,
const ClutterVertex *point,
ClutterVertex *vertex)
{
gfloat w;
CoglMatrix matrix;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (ancestor == NULL || CLUTTER_IS_ACTOR (ancestor));
g_return_if_fail (point != NULL);
g_return_if_fail (vertex != NULL);
*vertex = *point;
w = 1.0;
if (ancestor == NULL)
ancestor = _clutter_actor_get_stage_internal (self);
if (ancestor == NULL)
{
*vertex = *point;
return;
}
_clutter_actor_get_relative_transformation_matrix (self, ancestor, &matrix);
cogl_matrix_transform_point (&matrix, &vertex->x, &vertex->y, &vertex->z, &w);
}
static gboolean
_clutter_actor_fully_transform_vertices (ClutterActor *self,
const ClutterVertex *vertices_in,
ClutterVertex *vertices_out,
int n_vertices)
{
ClutterActor *stage;
CoglMatrix modelview;
CoglMatrix projection;
float viewport[4];
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
stage = _clutter_actor_get_stage_internal (self);
/* We really can't do anything meaningful in this case so don't try
* to do any transform */
if (stage == NULL)
return FALSE;
/* Note: we pass NULL as the ancestor because we don't just want the modelview
* that gets us to stage coordinates, we want to go all the way to eye
* coordinates */
_clutter_actor_apply_relative_transformation_matrix (self, NULL, &modelview);
/* Fetch the projection and viewport */
_clutter_stage_get_projection_matrix (CLUTTER_STAGE (stage), &projection);
_clutter_stage_get_viewport (CLUTTER_STAGE (stage),
&viewport[0],
&viewport[1],
&viewport[2],
&viewport[3]);
_clutter_util_fully_transform_vertices (&modelview,
&projection,
viewport,
vertices_in,
vertices_out,
n_vertices);
return TRUE;
}
/**
* clutter_actor_apply_transform_to_point:
* @self: A #ClutterActor
* @point: A point as #ClutterVertex
* @vertex: (out caller-allocates): The translated #ClutterVertex
*
* Transforms @point in coordinates relative to the actor
* into screen-relative coordinates with the current actor
* transformation (i.e. scale, rotation, etc)
*
* Since: 0.4
**/
void
clutter_actor_apply_transform_to_point (ClutterActor *self,
const ClutterVertex *point,
ClutterVertex *vertex)
2007-06-12 11:42:29 +00:00
{
g_return_if_fail (point != NULL);
g_return_if_fail (vertex != NULL);
_clutter_actor_fully_transform_vertices (self, point, vertex, 1);
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
}
/*
* _clutter_actor_get_relative_transformation_matrix:
* @self: The actor whose coordinate space you want to transform from.
* @ancestor: The ancestor actor whose coordinate space you want to transform too
* or %NULL if you want to transform all the way to eye coordinates.
* @matrix: A #CoglMatrix to store the transformation
*
* This gets a transformation @matrix that will transform coordinates from the
* coordinate space of @self into the coordinate space of @ancestor.
*
* For example if you need a matrix that can transform the local actor
* coordinates of @self into stage coordinates you would pass the actor's stage
* pointer as the @ancestor.
*
* If you pass %NULL then the transformation will take you all the way through
* to eye coordinates. This can be useful if you want to extract the entire
* modelview transform that Clutter applies before applying the projection
* transformation. If you want to explicitly set a modelview on a CoglFramebuffer
* using cogl_set_modelview_matrix() for example then you would want a matrix
* that transforms into eye coordinates.
*
* <note><para>This function explicitly initializes the given @matrix. If you just
* want clutter to multiply a relative transformation with an existing matrix
* you can use clutter_actor_apply_relative_transformation_matrix()
* instead.</para></note>
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
*
*/
/* XXX: We should consider caching the stage relative modelview along with
* the actor itself */
static void
_clutter_actor_get_relative_transformation_matrix (ClutterActor *self,
ClutterActor *ancestor,
CoglMatrix *matrix)
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
{
cogl_matrix_init_identity (matrix);
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
_clutter_actor_apply_relative_transformation_matrix (self, ancestor, matrix);
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
}
/* Project the given @box into stage window coordinates, writing the
* transformed vertices to @verts[]. */
static gboolean
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
_clutter_actor_transform_and_project_box (ClutterActor *self,
const ClutterActorBox *box,
ClutterVertex verts[])
{
ClutterVertex box_vertices[4];
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
box_vertices[0].x = box->x1;
box_vertices[0].y = box->y1;
box_vertices[0].z = 0;
box_vertices[1].x = box->x2;
box_vertices[1].y = box->y1;
box_vertices[1].z = 0;
box_vertices[2].x = box->x1;
box_vertices[2].y = box->y2;
box_vertices[2].z = 0;
box_vertices[3].x = box->x2;
box_vertices[3].y = box->y2;
box_vertices[3].z = 0;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
return
_clutter_actor_fully_transform_vertices (self, box_vertices, verts, 4);
}
/**
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* clutter_actor_get_allocation_vertices:
* @self: A #ClutterActor
2009-07-02 13:21:16 +00:00
* @ancestor: (allow-none): A #ClutterActor to calculate the vertices
* against, or %NULL to use the #ClutterStage
2010-09-08 15:41:47 +00:00
* @verts: (out) (array fixed-size=4) (element-type Clutter.Vertex): return
* location for an array of 4 #ClutterVertex in which to store the result
*
* Calculates the transformed coordinates of the four corners of the
* actor in the plane of @ancestor. The returned vertices relate to
* the #ClutterActorBox coordinates as follows:
* <itemizedlist>
* <listitem><para>@verts[0] contains (x1, y1)</para></listitem>
* <listitem><para>@verts[1] contains (x2, y1)</para></listitem>
* <listitem><para>@verts[2] contains (x1, y2)</para></listitem>
* <listitem><para>@verts[3] contains (x2, y2)</para></listitem>
* </itemizedlist>
*
* If @ancestor is %NULL the ancestor will be the #ClutterStage. In
* this case, the coordinates returned will be the coordinates on
* the stage before the projection is applied. This is different from
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* the behaviour of clutter_actor_get_abs_allocation_vertices().
*
* Since: 0.6
*/
void
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_allocation_vertices (ClutterActor *self,
ClutterActor *ancestor,
ClutterVertex verts[])
{
ClutterActorPrivate *priv;
ClutterActorBox box;
ClutterVertex vertices[4];
CoglMatrix modelview;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (ancestor == NULL || CLUTTER_IS_ACTOR (ancestor));
if (ancestor == NULL)
ancestor = _clutter_actor_get_stage_internal (self);
/* Fallback to a NOP transform if the actor isn't parented under a
* stage. */
if (ancestor == NULL)
ancestor = self;
priv = self->priv;
/* if the actor needs to be allocated we force a relayout, so that
* we will have valid values to use in the transformations */
if (priv->needs_allocation)
{
ClutterActor *stage = _clutter_actor_get_stage_internal (self);
if (stage)
_clutter_stage_maybe_relayout (stage);
else
{
box.x1 = box.y1 = 0;
/* The result isn't really meaningful in this case but at
* least try to do something *vaguely* reasonable... */
clutter_actor_get_size (self, &box.x2, &box.y2);
}
}
clutter_actor_get_allocation_box (self, &box);
vertices[0].x = box.x1;
vertices[0].y = box.y1;
vertices[0].z = 0;
vertices[1].x = box.x2;
vertices[1].y = box.y1;
vertices[1].z = 0;
vertices[2].x = box.x1;
vertices[2].y = box.y2;
vertices[2].z = 0;
vertices[3].x = box.x2;
vertices[3].y = box.y2;
vertices[3].z = 0;
_clutter_actor_get_relative_transformation_matrix (self, ancestor,
&modelview);
cogl_matrix_transform_points (&modelview,
3,
sizeof (ClutterVertex),
vertices,
sizeof (ClutterVertex),
vertices,
4);
}
/**
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* clutter_actor_get_abs_allocation_vertices:
* @self: A #ClutterActor
2009-06-10 13:54:42 +00:00
* @verts: (out) (array fixed-size=4): Pointer to a location of an array
* of 4 #ClutterVertex where to store the result.
*
* Calculates the transformed screen coordinates of the four corners of
* the actor; the returned vertices relate to the #ClutterActorBox
* coordinates as follows:
* <itemizedlist>
* <listitem><para>v[0] contains (x1, y1)</para></listitem>
* <listitem><para>v[1] contains (x2, y1)</para></listitem>
* <listitem><para>v[2] contains (x1, y2)</para></listitem>
* <listitem><para>v[3] contains (x2, y2)</para></listitem>
* </itemizedlist>
*
2007-06-12 11:42:29 +00:00
* Since: 0.4
*/
void
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_abs_allocation_vertices (ClutterActor *self,
ClutterVertex verts[])
{
ClutterActorPrivate *priv;
ClutterActorBox actor_space_allocation;
2007-03-22 Emmanuele Bassi <ebassi@openedhand.com> * clutter/clutter-private.h: Remove inclusion of backend-specific headers; update the main context object; add the declarations for the event queue functions. * clutter/clutter-backend.[ch]: Add the abstract ClutterBackend object, which holds backend-specific settings, the main stage, and the event queue. Every backend must implement a subclass of ClutterBackend and ClutterStage. * clutter/clutter-feature.c: Protect the GLX specific calls behing #ifdef HAVE_CLUTTER_GLX. * clutter/clutter-actor.c: * clutter/clutter-group.c: * clutter/clutter-clone-texture.c: Include GL/gl.h * clutter/clutter-event.[ch]: Update public API and implement the event queue private API; hold a reference on the event objects; move out the keysym-to-unicode table; add the new event types. * clutter/clutter-color.h: Include clutter-fixed.h * clutter/clutter-main.c: Update API; get the main stage from the backend object; process the event received from the queue; lock/unlock the main mutex if we have one; move the initialisation process sooner in the init sequence, in order to have the backend object when we check for options; call the backed vfuncs in the pre/post parse hooks. * clutter/clutter-stage.c: Make ClutterStage and abstract class, implemented by the backends. * clutter/clutter/glx/clutter-glx.h: * clutter/clutter/glx/clutter-backend-glx.[ch]: * clutter/clutter/glx/clutter-event-glx.c: * clutter/clutter/glx/clutter-stage-glx.[ch]: * clutter/clutter/glx/Makefile.am: Add the GLX backend. * clutter/clutter/egl/clutter-backend-egl.[ch]: * clutter/clutter/egl/clutter-event-egl.c: * clutter/clutter/egl/clutter-stage-egl.[ch]: * clutter/clutter/egl/Makefile.am: Add the stub for a EGL backend. * examples/*.c: Update for the new API.
2007-03-22 18:21:59 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
2007-06-12 11:42:29 +00:00
2007-03-22 Emmanuele Bassi <ebassi@openedhand.com> * clutter/clutter-private.h: Remove inclusion of backend-specific headers; update the main context object; add the declarations for the event queue functions. * clutter/clutter-backend.[ch]: Add the abstract ClutterBackend object, which holds backend-specific settings, the main stage, and the event queue. Every backend must implement a subclass of ClutterBackend and ClutterStage. * clutter/clutter-feature.c: Protect the GLX specific calls behing #ifdef HAVE_CLUTTER_GLX. * clutter/clutter-actor.c: * clutter/clutter-group.c: * clutter/clutter-clone-texture.c: Include GL/gl.h * clutter/clutter-event.[ch]: Update public API and implement the event queue private API; hold a reference on the event objects; move out the keysym-to-unicode table; add the new event types. * clutter/clutter-color.h: Include clutter-fixed.h * clutter/clutter-main.c: Update API; get the main stage from the backend object; process the event received from the queue; lock/unlock the main mutex if we have one; move the initialisation process sooner in the init sequence, in order to have the backend object when we check for options; call the backed vfuncs in the pre/post parse hooks. * clutter/clutter-stage.c: Make ClutterStage and abstract class, implemented by the backends. * clutter/clutter/glx/clutter-glx.h: * clutter/clutter/glx/clutter-backend-glx.[ch]: * clutter/clutter/glx/clutter-event-glx.c: * clutter/clutter/glx/clutter-stage-glx.[ch]: * clutter/clutter/glx/Makefile.am: Add the GLX backend. * clutter/clutter/egl/clutter-backend-egl.[ch]: * clutter/clutter/egl/clutter-event-egl.c: * clutter/clutter/egl/clutter-stage-egl.[ch]: * clutter/clutter/egl/Makefile.am: Add the stub for a EGL backend. * examples/*.c: Update for the new API.
2007-03-22 18:21:59 +00:00
priv = self->priv;
/* if the actor needs to be allocated we force a relayout, so that
* the actor allocation box will be valid for
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
* _clutter_actor_transform_and_project_box()
*/
if (priv->needs_allocation)
{
ClutterActor *stage = _clutter_actor_get_stage_internal (self);
/* There's nothing meaningful we can do now */
if (!stage)
return;
_clutter_stage_maybe_relayout (stage);
}
2007-06-12 11:42:29 +00:00
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
/* NB: _clutter_actor_transform_and_project_box expects a box in the actor's
* own coordinate space... */
actor_space_allocation.x1 = 0;
actor_space_allocation.y1 = 0;
actor_space_allocation.x2 = priv->allocation.x2 - priv->allocation.x1;
actor_space_allocation.y2 = priv->allocation.y2 - priv->allocation.y1;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
_clutter_actor_transform_and_project_box (self,
&actor_space_allocation,
verts);
2007-06-12 11:42:29 +00:00
}
static void
[actor] Allow changing the transformations matrix Currently, the transformation matrix for an actor is constructed from scenegraph-related accessors. An actor, though, can call COGL API to add new transformations inside the paint() implementation, for instance: static void my_foo_paint (ClutterActor *a) { ... cogl_translate (-scroll_x, -scroll_y, 0); ... } Unfortunately these transformations will be completely ignored by the scenegraph machinery; for instance, getting the actor-relative coordinates from event coordinates is going to break badly because of this. In order to make the scenegraph aware of the potential of additional transformations, we need a ::apply_transform() virtual function. This vfunc will pass a CoglMatrix which can be used to apply additional operations: static void my_foo_apply_transform (ClutterActor *a, CoglMatrix *m) { CLUTTER_ACTOR_CLASS (my_foo_parent_class)->apply_transform (a, m); ... cogl_matrix_translate (m, -scroll_x, -scroll_y, 0); ... } The ::paint() implementation will be called with the actor already using the newly applied transformation matrix, as expected: static void my_foo_paint (ClutterActor *a) { ... } The ::apply_transform() implementations *must* chain up, so that the various transformations of each class are preserved. The default implementation inside ClutterActor applies all the transformations defined by the scenegraph-related accessors. Actors performing transformations inside the paint() function will continue to work as previously.
2009-07-01 12:59:13 +00:00
clutter_actor_real_apply_transform (ClutterActor *self,
CoglMatrix *matrix)
2007-06-12 11:42:29 +00:00
{
ClutterActorPrivate *priv = self->priv;
if (!priv->transform_valid)
{
CoglMatrix *transform = &priv->transform;
const ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info_or_defaults (self);
cogl_matrix_init_identity (transform);
cogl_matrix_translate (transform,
priv->allocation.x1,
priv->allocation.y1,
0.0);
if (priv->z)
cogl_matrix_translate (transform, 0, 0, priv->z);
/*
* because the rotation involves translations, we must scale
* before applying the rotations (if we apply the scale after
* the rotations, the translations included in the rotation are
* not scaled and so the entire object will move on the screen
* as a result of rotating it).
*/
if (info->scale_x != 1.0 || info->scale_y != 1.0)
{
TRANSFORM_ABOUT_ANCHOR_COORD (self, transform,
&info->scale_center,
cogl_matrix_scale (transform,
info->scale_x,
info->scale_y,
1.0));
}
if (info->rz_angle)
TRANSFORM_ABOUT_ANCHOR_COORD (self, transform,
&info->rz_center,
cogl_matrix_rotate (transform,
info->rz_angle,
0, 0, 1.0));
if (info->ry_angle)
TRANSFORM_ABOUT_ANCHOR_COORD (self, transform,
&info->ry_center,
cogl_matrix_rotate (transform,
info->ry_angle,
0, 1.0, 0));
if (info->rx_angle)
TRANSFORM_ABOUT_ANCHOR_COORD (self, transform,
&info->rx_center,
cogl_matrix_rotate (transform,
info->rx_angle,
1.0, 0, 0));
if (!clutter_anchor_coord_is_zero (&info->anchor))
{
gfloat x, y, z;
clutter_anchor_coord_get_units (self, &info->anchor, &x, &y, &z);
cogl_matrix_translate (transform, -x, -y, -z);
}
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
priv->transform_valid = TRUE;
}
cogl_matrix_multiply (matrix, matrix, &priv->transform);
2007-06-12 11:42:29 +00:00
}
/* Applies the transforms associated with this actor to the given
* matrix. */
void
_clutter_actor_apply_modelview_transform (ClutterActor *self,
CoglMatrix *matrix)
[actor] Allow changing the transformations matrix Currently, the transformation matrix for an actor is constructed from scenegraph-related accessors. An actor, though, can call COGL API to add new transformations inside the paint() implementation, for instance: static void my_foo_paint (ClutterActor *a) { ... cogl_translate (-scroll_x, -scroll_y, 0); ... } Unfortunately these transformations will be completely ignored by the scenegraph machinery; for instance, getting the actor-relative coordinates from event coordinates is going to break badly because of this. In order to make the scenegraph aware of the potential of additional transformations, we need a ::apply_transform() virtual function. This vfunc will pass a CoglMatrix which can be used to apply additional operations: static void my_foo_apply_transform (ClutterActor *a, CoglMatrix *m) { CLUTTER_ACTOR_CLASS (my_foo_parent_class)->apply_transform (a, m); ... cogl_matrix_translate (m, -scroll_x, -scroll_y, 0); ... } The ::paint() implementation will be called with the actor already using the newly applied transformation matrix, as expected: static void my_foo_paint (ClutterActor *a) { ... } The ::apply_transform() implementations *must* chain up, so that the various transformations of each class are preserved. The default implementation inside ClutterActor applies all the transformations defined by the scenegraph-related accessors. Actors performing transformations inside the paint() function will continue to work as previously.
2009-07-01 12:59:13 +00:00
{
CLUTTER_ACTOR_GET_CLASS (self)->apply_transform (self, matrix);
[actor] Allow changing the transformations matrix Currently, the transformation matrix for an actor is constructed from scenegraph-related accessors. An actor, though, can call COGL API to add new transformations inside the paint() implementation, for instance: static void my_foo_paint (ClutterActor *a) { ... cogl_translate (-scroll_x, -scroll_y, 0); ... } Unfortunately these transformations will be completely ignored by the scenegraph machinery; for instance, getting the actor-relative coordinates from event coordinates is going to break badly because of this. In order to make the scenegraph aware of the potential of additional transformations, we need a ::apply_transform() virtual function. This vfunc will pass a CoglMatrix which can be used to apply additional operations: static void my_foo_apply_transform (ClutterActor *a, CoglMatrix *m) { CLUTTER_ACTOR_CLASS (my_foo_parent_class)->apply_transform (a, m); ... cogl_matrix_translate (m, -scroll_x, -scroll_y, 0); ... } The ::paint() implementation will be called with the actor already using the newly applied transformation matrix, as expected: static void my_foo_paint (ClutterActor *a) { ... } The ::apply_transform() implementations *must* chain up, so that the various transformations of each class are preserved. The default implementation inside ClutterActor applies all the transformations defined by the scenegraph-related accessors. Actors performing transformations inside the paint() function will continue to work as previously.
2009-07-01 12:59:13 +00:00
}
/*
* clutter_actor_apply_relative_transformation_matrix:
* @self: The actor whose coordinate space you want to transform from.
* @ancestor: The ancestor actor whose coordinate space you want to transform too
* or %NULL if you want to transform all the way to eye coordinates.
* @matrix: A #CoglMatrix to apply the transformation too.
*
* This multiplies a transform with @matrix that will transform coordinates
* from the coordinate space of @self into the coordinate space of @ancestor.
*
* For example if you need a matrix that can transform the local actor
* coordinates of @self into stage coordinates you would pass the actor's stage
* pointer as the @ancestor.
*
* If you pass %NULL then the transformation will take you all the way through
* to eye coordinates. This can be useful if you want to extract the entire
* modelview transform that Clutter applies before applying the projection
* transformation. If you want to explicitly set a modelview on a CoglFramebuffer
* using cogl_set_modelview_matrix() for example then you would want a matrix
* that transforms into eye coordinates.
*
* <note>This function doesn't initialize the given @matrix, it simply
* multiplies the requested transformation matrix with the existing contents of
* @matrix. You can use cogl_matrix_init_identity() to initialize the @matrix
* before calling this function, or you can use
* clutter_actor_get_relative_transformation_matrix() instead.</note>
*/
void
_clutter_actor_apply_relative_transformation_matrix (ClutterActor *self,
ClutterActor *ancestor,
CoglMatrix *matrix)
2007-06-12 11:42:29 +00:00
{
ClutterActor *parent;
/* Note we terminate before ever calling stage->apply_transform()
* since that would conceptually be relative to the underlying
* window OpenGL coordinates so we'd need a special @ancestor
* value to represent the fake parent of the stage. */
if (self == ancestor)
return;
parent = clutter_actor_get_parent (self);
if (parent != NULL)
_clutter_actor_apply_relative_transformation_matrix (parent, ancestor,
matrix);
_clutter_actor_apply_modelview_transform (self, matrix);
2007-06-12 11:42:29 +00:00
}
static void
_clutter_actor_draw_paint_volume_full (ClutterActor *self,
ClutterPaintVolume *pv,
const char *label,
const CoglColor *color)
{
static CoglPipeline *outline = NULL;
CoglPrimitive *prim;
ClutterVertex line_ends[12 * 2];
int n_vertices;
CoglContext *ctx =
clutter_backend_get_cogl_context (clutter_get_default_backend ());
/* XXX: at some point we'll query this from the stage but we can't
* do that until the osx backend uses Cogl natively. */
CoglFramebuffer *fb = cogl_get_draw_framebuffer ();
if (outline == NULL)
outline = cogl_pipeline_new ();
_clutter_paint_volume_complete (pv);
n_vertices = pv->is_2d ? 4 * 2 : 12 * 2;
/* Front face */
line_ends[0] = pv->vertices[0]; line_ends[1] = pv->vertices[1];
line_ends[2] = pv->vertices[1]; line_ends[3] = pv->vertices[2];
line_ends[4] = pv->vertices[2]; line_ends[5] = pv->vertices[3];
line_ends[6] = pv->vertices[3]; line_ends[7] = pv->vertices[0];
if (!pv->is_2d)
{
/* Back face */
line_ends[8] = pv->vertices[4]; line_ends[9] = pv->vertices[5];
line_ends[10] = pv->vertices[5]; line_ends[11] = pv->vertices[6];
line_ends[12] = pv->vertices[6]; line_ends[13] = pv->vertices[7];
line_ends[14] = pv->vertices[7]; line_ends[15] = pv->vertices[4];
/* Lines connecting front face to back face */
line_ends[16] = pv->vertices[0]; line_ends[17] = pv->vertices[4];
line_ends[18] = pv->vertices[1]; line_ends[19] = pv->vertices[5];
line_ends[20] = pv->vertices[2]; line_ends[21] = pv->vertices[6];
line_ends[22] = pv->vertices[3]; line_ends[23] = pv->vertices[7];
}
prim = cogl_primitive_new_p3 (ctx, COGL_VERTICES_MODE_LINES, n_vertices,
(CoglVertexP3 *)line_ends);
cogl_pipeline_set_color (outline, color);
cogl_framebuffer_draw_primitive (fb, outline, prim);
cogl_object_unref (prim);
if (label)
{
PangoLayout *layout;
layout = pango_layout_new (clutter_actor_get_pango_context (self));
pango_layout_set_text (layout, label, -1);
cogl_pango_render_layout (layout,
pv->vertices[0].x,
pv->vertices[0].y,
color,
0);
g_object_unref (layout);
}
}
static void
_clutter_actor_draw_paint_volume (ClutterActor *self)
{
ClutterPaintVolume *pv;
CoglColor color;
pv = _clutter_actor_get_paint_volume_mutable (self);
if (!pv)
{
gfloat width, height;
ClutterPaintVolume fake_pv;
ClutterActor *stage = _clutter_actor_get_stage_internal (self);
_clutter_paint_volume_init_static (&fake_pv, stage);
clutter_actor_get_size (self, &width, &height);
clutter_paint_volume_set_width (&fake_pv, width);
clutter_paint_volume_set_height (&fake_pv, height);
cogl_color_init_from_4f (&color, 0, 0, 1, 1);
_clutter_actor_draw_paint_volume_full (self, &fake_pv,
_clutter_actor_get_debug_name (self),
&color);
clutter_paint_volume_free (&fake_pv);
}
else
{
cogl_color_init_from_4f (&color, 0, 1, 0, 1);
_clutter_actor_draw_paint_volume_full (self, pv,
_clutter_actor_get_debug_name (self),
&color);
}
}
static void
_clutter_actor_paint_cull_result (ClutterActor *self,
gboolean success,
ClutterCullResult result)
{
ClutterPaintVolume *pv;
CoglColor color;
if (success)
{
if (result == CLUTTER_CULL_RESULT_IN)
cogl_color_init_from_4f (&color, 0, 1, 0, 1);
else if (result == CLUTTER_CULL_RESULT_OUT)
cogl_color_init_from_4f (&color, 0, 0, 1, 1);
else
cogl_color_init_from_4f (&color, 0, 1, 1, 1);
}
else
cogl_color_init_from_4f (&color, 1, 1, 1, 1);
if (success && (pv = _clutter_actor_get_paint_volume_mutable (self)))
_clutter_actor_draw_paint_volume_full (self, pv,
_clutter_actor_get_debug_name (self),
&color);
else
{
PangoLayout *layout;
char *label =
g_strdup_printf ("CULL FAILURE: %s", _clutter_actor_get_debug_name (self));
cogl_color_init_from_4f (&color, 1, 1, 1, 1);
cogl_set_source_color (&color);
layout = pango_layout_new (clutter_actor_get_pango_context (self));
pango_layout_set_text (layout, label, -1);
cogl_pango_render_layout (layout,
0,
0,
&color,
0);
g_free (label);
g_object_unref (layout);
}
}
static int clone_paint_level = 0;
void
_clutter_actor_push_clone_paint (void)
{
clone_paint_level++;
}
void
_clutter_actor_pop_clone_paint (void)
{
clone_paint_level--;
}
static gboolean
in_clone_paint (void)
{
return clone_paint_level > 0;
}
/* Returns TRUE if the actor can be ignored */
/* FIXME: we should return a ClutterCullResult, and
* clutter_actor_paint should understand that a CLUTTER_CULL_RESULT_IN
* means there's no point in trying to cull descendants of the current
* node. */
static gboolean
cull_actor (ClutterActor *self, ClutterCullResult *result_out)
{
ClutterActorPrivate *priv = self->priv;
ClutterActor *stage;
const ClutterPlane *stage_clip;
if (!priv->last_paint_volume_valid)
{
CLUTTER_NOTE (CLIPPING, "Bail from cull_actor without culling (%s): "
"->last_paint_volume_valid == FALSE",
_clutter_actor_get_debug_name (self));
return FALSE;
}
if (G_UNLIKELY (clutter_paint_debug_flags & CLUTTER_DEBUG_DISABLE_CULLING))
return FALSE;
stage = _clutter_actor_get_stage_internal (self);
stage_clip = _clutter_stage_get_clip (CLUTTER_STAGE (stage));
if (G_UNLIKELY (!stage_clip))
{
CLUTTER_NOTE (CLIPPING, "Bail from cull_actor without culling (%s): "
"No stage clip set",
_clutter_actor_get_debug_name (self));
return FALSE;
}
if (cogl_get_draw_framebuffer () !=
_clutter_stage_get_active_framebuffer (CLUTTER_STAGE (stage)))
{
CLUTTER_NOTE (CLIPPING, "Bail from cull_actor without culling (%s): "
"Current framebuffer doesn't correspond to stage",
_clutter_actor_get_debug_name (self));
return FALSE;
}
*result_out =
_clutter_paint_volume_cull (&priv->last_paint_volume, stage_clip);
return TRUE;
}
static void
_clutter_actor_update_last_paint_volume (ClutterActor *self)
{
ClutterActorPrivate *priv = self->priv;
const ClutterPaintVolume *pv;
if (priv->last_paint_volume_valid)
{
clutter_paint_volume_free (&priv->last_paint_volume);
priv->last_paint_volume_valid = FALSE;
}
pv = clutter_actor_get_paint_volume (self);
if (!pv)
{
CLUTTER_NOTE (CLIPPING, "Bail from update_last_paint_volume (%s): "
"Actor failed to report a paint volume",
_clutter_actor_get_debug_name (self));
return;
}
_clutter_paint_volume_copy_static (pv, &priv->last_paint_volume);
_clutter_paint_volume_transform_relative (&priv->last_paint_volume,
NULL); /* eye coordinates */
priv->last_paint_volume_valid = TRUE;
}
static inline gboolean
actor_has_shader_data (ClutterActor *self)
{
return g_object_get_qdata (G_OBJECT (self), quark_shader_data) != NULL;
}
guint32
_clutter_actor_get_pick_id (ClutterActor *self)
{
if (self->priv->pick_id < 0)
return 0;
return self->priv->pick_id;
}
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
/* This is the same as clutter_actor_add_effect except that it doesn't
queue a redraw and it doesn't notify on the effect property */
static void
_clutter_actor_add_effect_internal (ClutterActor *self,
ClutterEffect *effect)
{
ClutterActorPrivate *priv = self->priv;
if (priv->effects == NULL)
{
priv->effects = g_object_new (CLUTTER_TYPE_META_GROUP, NULL);
priv->effects->actor = self;
}
_clutter_meta_group_add_meta (priv->effects, CLUTTER_ACTOR_META (effect));
}
/* This is the same as clutter_actor_remove_effect except that it doesn't
queue a redraw and it doesn't notify on the effect property */
static void
_clutter_actor_remove_effect_internal (ClutterActor *self,
ClutterEffect *effect)
{
ClutterActorPrivate *priv = self->priv;
if (priv->effects == NULL)
return;
_clutter_meta_group_remove_meta (priv->effects, CLUTTER_ACTOR_META (effect));
}
static gboolean
needs_flatten_effect (ClutterActor *self)
{
ClutterActorPrivate *priv = self->priv;
if (G_UNLIKELY (clutter_paint_debug_flags &
CLUTTER_DEBUG_DISABLE_OFFSCREEN_REDIRECT))
return FALSE;
if (priv->offscreen_redirect & CLUTTER_OFFSCREEN_REDIRECT_ALWAYS)
return TRUE;
else if (priv->offscreen_redirect & CLUTTER_OFFSCREEN_REDIRECT_AUTOMATIC_FOR_OPACITY)
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
{
if (clutter_actor_get_paint_opacity (self) < 255 &&
clutter_actor_has_overlaps (self))
return TRUE;
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
}
return FALSE;
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
}
static void
add_or_remove_flatten_effect (ClutterActor *self)
{
ClutterActorPrivate *priv = self->priv;
/* Add or remove the flatten effect depending on the
offscreen-redirect property. */
if (needs_flatten_effect (self))
{
if (priv->flatten_effect == NULL)
{
ClutterActorMeta *actor_meta;
gint priority;
priv->flatten_effect = _clutter_flatten_effect_new ();
/* Keep a reference to the effect so that we can queue
redraws from it */
g_object_ref_sink (priv->flatten_effect);
/* Set the priority of the effect to high so that it will
always be applied to the actor first. It uses an internal
priority so that it won't be visible to applications */
actor_meta = CLUTTER_ACTOR_META (priv->flatten_effect);
priority = CLUTTER_ACTOR_META_PRIORITY_INTERNAL_HIGH;
_clutter_actor_meta_set_priority (actor_meta, priority);
/* This will add the effect without queueing a redraw */
_clutter_actor_add_effect_internal (self, priv->flatten_effect);
}
}
else
{
if (priv->flatten_effect != NULL)
{
/* Destroy the effect so that it will lose its fbo cache of
the actor */
_clutter_actor_remove_effect_internal (self, priv->flatten_effect);
g_object_unref (priv->flatten_effect);
priv->flatten_effect = NULL;
}
}
}
static void
clutter_actor_real_paint (ClutterActor *actor)
{
ClutterActorPrivate *priv = actor->priv;
ClutterActor *iter;
/* paint the background color, if set */
if (priv->bg_color_set)
{
float width, height;
guint8 real_alpha;
clutter_actor_box_get_size (&priv->allocation, &width, &height);
real_alpha = clutter_actor_get_paint_opacity_internal (actor)
* priv->bg_color.alpha
/ 255;
cogl_set_source_color4ub (priv->bg_color.red,
priv->bg_color.green,
priv->bg_color.blue,
real_alpha);
cogl_rectangle (0, 0, width, height);
}
for (iter = priv->first_child;
iter != NULL;
iter = iter->priv->next_sibling)
{
CLUTTER_NOTE (PAINT, "Painting %s, child of %s, at { %.2f, %.2f - %.2f x %.2f }",
2011-12-20 14:47:04 +00:00
_clutter_actor_get_debug_name (iter),
_clutter_actor_get_debug_name (actor),
2011-12-20 14:47:04 +00:00
iter->priv->allocation.x1,
iter->priv->allocation.y1,
iter->priv->allocation.x2 - iter->priv->allocation.x1,
iter->priv->allocation.y2 - iter->priv->allocation.y1);
clutter_actor_paint (iter);
}
}
2007-06-12 11:42:29 +00:00
/**
* clutter_actor_paint:
* @self: A #ClutterActor
*
* Renders the actor to display.
*
* This function should not be called directly by applications.
* Call clutter_actor_queue_redraw() to queue paints, instead.
*
2010-08-18 17:20:04 +00:00
* This function is context-aware, and will either cause a
* regular paint or a pick paint.
*
* This function will emit the #ClutterActor::paint signal or
* the #ClutterActor::pick signal, depending on the context.
*
* This function does not paint the actor if the actor is set to 0,
* unless it is performing a pick paint.
*/
2007-06-12 11:42:29 +00:00
void
clutter_actor_paint (ClutterActor *self)
{
ClutterActorPrivate *priv;
ClutterPickMode pick_mode;
gboolean clip_set = FALSE;
gboolean shader_applied = FALSE;
2009-03-30 15:41:02 +00:00
CLUTTER_STATIC_COUNTER (actor_paint_counter,
"Actor real-paint counter",
"Increments each time any actor is painted",
0 /* no application private data */);
CLUTTER_STATIC_COUNTER (actor_pick_counter,
"Actor pick-paint counter",
"Increments each time any actor is painted "
"for picking",
0 /* no application private data */);
2007-06-12 11:42:29 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (CLUTTER_ACTOR_IN_DESTRUCTION (self))
return;
2007-06-12 11:42:29 +00:00
priv = self->priv;
pick_mode = _clutter_context_get_pick_mode ();
if (pick_mode == CLUTTER_PICK_NONE)
priv->propagated_one_redraw = FALSE;
/* It's an important optimization that we consider painting of
* actors with 0 opacity to be a NOP... */
if (pick_mode == CLUTTER_PICK_NONE &&
/* ignore top-levels, since they might be transparent */
!CLUTTER_ACTOR_IS_TOPLEVEL (self) &&
/* Use the override opacity if its been set */
((priv->opacity_override >= 0) ?
priv->opacity_override : priv->opacity) == 0)
return;
2007-06-12 11:42:29 +00:00
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* if we aren't paintable (not in a toplevel with all
* parents paintable) then do nothing.
*/
if (!CLUTTER_ACTOR_IS_MAPPED (self))
return;
2007-06-12 11:42:29 +00:00
/* mark that we are in the paint process */
CLUTTER_SET_PRIVATE_FLAGS (self, CLUTTER_IN_PAINT);
2007-06-12 11:42:29 +00:00
cogl_push_matrix();
if (priv->enable_model_view_transform)
{
CoglMatrix matrix;
/* XXX: It could be better to cache the modelview with the actor
* instead of progressively building up the transformations on
* the matrix stack every time we paint. */
cogl_get_modelview_matrix (&matrix);
_clutter_actor_apply_modelview_transform (self, &matrix);
#ifdef CLUTTER_ENABLE_DEBUG
/* Catch when out-of-band transforms have been made by actors not as part
* of an apply_transform vfunc... */
if (G_UNLIKELY (clutter_debug_flags & CLUTTER_DEBUG_OOB_TRANSFORMS))
{
CoglMatrix expected_matrix;
_clutter_actor_get_relative_transformation_matrix (self, NULL,
&expected_matrix);
if (!cogl_matrix_equal (&matrix, &expected_matrix))
{
GString *buf = g_string_sized_new (1024);
ClutterActor *parent;
parent = self;
while (parent != NULL)
{
g_string_append (buf, _clutter_actor_get_debug_name (parent));
if (parent->priv->parent != NULL)
g_string_append (buf, "->");
parent = parent->priv->parent;
}
g_warning ("Unexpected transform found when painting actor "
"\"%s\". This will be caused by one of the actor's "
"ancestors (%s) using the Cogl API directly to transform "
"children instead of using ::apply_transform().",
_clutter_actor_get_debug_name (self),
buf->str);
g_string_free (buf, TRUE);
}
}
#endif /* CLUTTER_ENABLE_DEBUG */
cogl_set_modelview_matrix (&matrix);
}
if (priv->has_clip)
{
cogl_clip_push_rectangle (priv->clip.x,
priv->clip.y,
priv->clip.x + priv->clip.width,
priv->clip.y + priv->clip.height);
clip_set = TRUE;
}
else if (priv->clip_to_allocation)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat width, height;
width = priv->allocation.x2 - priv->allocation.x1;
height = priv->allocation.y2 - priv->allocation.y1;
cogl_clip_push_rectangle (0, 0, width, height);
clip_set = TRUE;
}
if (pick_mode == CLUTTER_PICK_NONE)
{
CLUTTER_COUNTER_INC (_clutter_uprof_context, actor_paint_counter);
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
/* We check whether we need to add the flatten effect before
each paint so that we can avoid having a mechanism for
applications to notify when the value of the
has_overlaps virtual changes. */
add_or_remove_flatten_effect (self);
}
else
CLUTTER_COUNTER_INC (_clutter_uprof_context, actor_pick_counter);
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
/* We save the current paint volume so that the next time the
* actor queues a redraw we can constrain the redraw to just
* cover the union of the new bounding box and the old.
*
* We also fetch the current paint volume to perform culling so
* we can avoid painting actors outside the current clip region.
*
* If we are painting inside a clone, we should neither update
* the paint volume or use it to cull painting, since the paint
* box represents the location of the source actor on the
* screen.
*
* XXX: We are starting to do a lot of vertex transforms on
* the CPU in a typical paint, so at some point we should
* audit these and consider caching some things.
*
* NB: We don't perform culling while picking at this point because
* clutter-stage.c doesn't setup the clipping planes appropriately.
*
* NB: We don't want to update the last-paint-volume during picking
* because the last-paint-volume is used to determine the old screen
* space location of an actor that has moved so we can know the
* minimal region to redraw to clear an old view of the actor. If we
* update this during picking then by the time we come around to
* paint then the last-paint-volume would likely represent the new
* actor position not the old.
*/
if (!in_clone_paint () && pick_mode == CLUTTER_PICK_NONE)
{
gboolean success;
/* annoyingly gcc warns if uninitialized even though
* the initialization is redundant :-( */
ClutterCullResult result = CLUTTER_CULL_RESULT_IN;
if (G_LIKELY ((clutter_paint_debug_flags &
(CLUTTER_DEBUG_DISABLE_CULLING |
CLUTTER_DEBUG_DISABLE_CLIPPED_REDRAWS)) !=
(CLUTTER_DEBUG_DISABLE_CULLING |
CLUTTER_DEBUG_DISABLE_CLIPPED_REDRAWS)))
_clutter_actor_update_last_paint_volume (self);
success = cull_actor (self, &result);
if (G_UNLIKELY (clutter_paint_debug_flags & CLUTTER_DEBUG_REDRAWS))
_clutter_actor_paint_cull_result (self, success, result);
else if (result == CLUTTER_CULL_RESULT_OUT && success)
goto done;
}
if (priv->effects == NULL)
{
if (pick_mode == CLUTTER_PICK_NONE &&
actor_has_shader_data (self))
{
_clutter_actor_shader_pre_paint (self, FALSE);
shader_applied = TRUE;
}
priv->next_effect_to_paint = NULL;
}
else
priv->next_effect_to_paint =
_clutter_meta_group_peek_metas (priv->effects);
clutter_actor_continue_paint (self);
2009-03-30 15:41:02 +00:00
if (shader_applied)
_clutter_actor_shader_post_paint (self);
if (G_UNLIKELY (clutter_paint_debug_flags & CLUTTER_DEBUG_PAINT_VOLUMES &&
pick_mode == CLUTTER_PICK_NONE))
_clutter_actor_draw_paint_volume (self);
done:
/* If we make it here then the actor has run through a complete
paint run including all the effects so it's no longer dirty */
if (pick_mode == CLUTTER_PICK_NONE)
priv->is_dirty = FALSE;
if (clip_set)
cogl_clip_pop();
cogl_pop_matrix();
/* paint sequence complete */
CLUTTER_UNSET_PRIVATE_FLAGS (self, CLUTTER_IN_PAINT);
}
/**
* clutter_actor_continue_paint:
* @self: A #ClutterActor
*
* Run the next stage of the paint sequence. This function should only
* be called within the implementation of the run virtual of a
* #ClutterEffect. It will cause the run method of the next effect to
* be applied, or it will paint the actual actor if the current effect
* is the last effect in the chain.
*
* Since: 1.8
*/
void
clutter_actor_continue_paint (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
/* This should only be called from with in the run implementation
of a ClutterEffect */
g_return_if_fail (CLUTTER_ACTOR_IN_PAINT (self));
priv = self->priv;
/* Skip any effects that are disabled */
while (priv->next_effect_to_paint &&
!clutter_actor_meta_get_enabled (priv->next_effect_to_paint->data))
priv->next_effect_to_paint = priv->next_effect_to_paint->next;
/* If this has come from the last effect then we'll just paint the
actual actor */
if (priv->next_effect_to_paint == NULL)
{
if (_clutter_context_get_pick_mode () == CLUTTER_PICK_NONE)
{
g_signal_emit (self, actor_signals[PAINT], 0);
}
else
{
ClutterColor col = { 0, };
_clutter_id_to_color (_clutter_actor_get_pick_id (self), &col);
/* Actor will then paint silhouette of itself in supplied
* color. See clutter_stage_get_actor_at_pos() for where
* picking is enabled.
*/
g_signal_emit (self, actor_signals[PICK], 0, &col);
}
}
else
{
ClutterEffect *old_current_effect;
ClutterEffectPaintFlags run_flags = 0;
/* Cache the current effect so that we can put it back before
returning */
old_current_effect = priv->current_effect;
priv->current_effect = priv->next_effect_to_paint->data;
priv->next_effect_to_paint = priv->next_effect_to_paint->next;
if (_clutter_context_get_pick_mode () == CLUTTER_PICK_NONE)
{
if (priv->is_dirty)
{
/* If there's an effect queued with this redraw then all
effects up to that one will be considered dirty. It
is expected the queued effect will paint the cached
image and not call clutter_actor_continue_paint again
(although it should work ok if it does) */
if (priv->effect_to_redraw == NULL ||
priv->current_effect != priv->effect_to_redraw)
run_flags |= CLUTTER_EFFECT_PAINT_ACTOR_DIRTY;
}
_clutter_effect_paint (priv->current_effect, run_flags);
}
else
{
/* We can't determine when an actor has been modified since
its last pick so lets just assume it has always been
modified */
run_flags |= CLUTTER_EFFECT_PAINT_ACTOR_DIRTY;
_clutter_effect_pick (priv->current_effect, run_flags);
}
priv->current_effect = old_current_effect;
}
}
static ClutterActorTraverseVisitFlags
invalidate_queue_redraw_entry (ClutterActor *self,
int depth,
gpointer user_data)
{
ClutterActorPrivate *priv = self->priv;
if (priv->queue_redraw_entry != NULL)
{
_clutter_stage_queue_redraw_entry_invalidate (priv->queue_redraw_entry);
priv->queue_redraw_entry = NULL;
}
return CLUTTER_ACTOR_TRAVERSE_VISIT_CONTINUE;
}
static inline void
remove_child (ClutterActor *self,
ClutterActor *child)
{
ClutterActor *prev_sibling, *next_sibling;
prev_sibling = child->priv->prev_sibling;
next_sibling = child->priv->next_sibling;
if (prev_sibling != NULL)
prev_sibling->priv->next_sibling = next_sibling;
if (next_sibling != NULL)
next_sibling->priv->prev_sibling = prev_sibling;
if (self->priv->first_child == child)
self->priv->first_child = next_sibling;
if (self->priv->last_child == child)
self->priv->last_child = prev_sibling;
child->priv->parent = NULL;
child->priv->prev_sibling = NULL;
child->priv->next_sibling = NULL;
}
typedef enum {
REMOVE_CHILD_DESTROY_META = 1 << 0,
REMOVE_CHILD_EMIT_PARENT_SET = 1 << 1,
REMOVE_CHILD_EMIT_ACTOR_REMOVED = 1 << 2,
REMOVE_CHILD_CHECK_STATE = 1 << 3,
REMOVE_CHILD_FLUSH_QUEUE = 1 << 4,
REMOVE_CHILD_NOTIFY_FIRST_LAST = 1 << 5,
/* default flags for public API */
REMOVE_CHILD_DEFAULT_FLAGS = REMOVE_CHILD_DESTROY_META |
REMOVE_CHILD_EMIT_PARENT_SET |
REMOVE_CHILD_EMIT_ACTOR_REMOVED |
REMOVE_CHILD_CHECK_STATE |
REMOVE_CHILD_FLUSH_QUEUE |
REMOVE_CHILD_NOTIFY_FIRST_LAST,
/* flags for legacy/deprecated API */
REMOVE_CHILD_LEGACY_FLAGS = REMOVE_CHILD_CHECK_STATE |
REMOVE_CHILD_FLUSH_QUEUE |
REMOVE_CHILD_EMIT_PARENT_SET |
REMOVE_CHILD_NOTIFY_FIRST_LAST
} ClutterActorRemoveChildFlags;
/*< private >
* clutter_actor_remove_child_internal:
* @self: a #ClutterActor
* @child: the child of @self that has to be removed
* @flags: control the removal operations
*
* Removes @child from the list of children of @self.
*/
static void
clutter_actor_remove_child_internal (ClutterActor *self,
ClutterActor *child,
ClutterActorRemoveChildFlags flags)
{
ClutterActor *old_first, *old_last;
gboolean destroy_meta, emit_parent_set, emit_actor_removed, check_state;
gboolean flush_queue;
gboolean notify_first_last;
gboolean was_mapped;
destroy_meta = (flags & REMOVE_CHILD_DESTROY_META) != 0;
emit_parent_set = (flags & REMOVE_CHILD_EMIT_PARENT_SET) != 0;
emit_actor_removed = (flags & REMOVE_CHILD_EMIT_ACTOR_REMOVED) != 0;
check_state = (flags & REMOVE_CHILD_CHECK_STATE) != 0;
flush_queue = (flags & REMOVE_CHILD_FLUSH_QUEUE) != 0;
notify_first_last = (flags & REMOVE_CHILD_NOTIFY_FIRST_LAST) != 0;
if (destroy_meta)
clutter_container_destroy_child_meta (CLUTTER_CONTAINER (self), child);
if (check_state)
{
was_mapped = CLUTTER_ACTOR_IS_MAPPED (child);
/* we need to unrealize *before* we set parent_actor to NULL,
* because in an unrealize method actors are dissociating from the
* stage, which means they need to be able to
* clutter_actor_get_stage().
*
* yhis should unmap and unrealize, unless we're reparenting.
*/
clutter_actor_update_map_state (child, MAP_STATE_MAKE_UNREALIZED);
}
else
was_mapped = FALSE;
if (flush_queue)
{
/* We take this opportunity to invalidate any queue redraw entry
* associated with the actor and descendants since we won't be able to
* determine the appropriate stage after this.
*
* we do this after we updated the mapped state because actors might
* end up queueing redraws inside their mapped/unmapped virtual
* functions, and if we invalidate the redraw entry we could end up
* with an inconsistent state and weird memory corruption. see
* bugs:
*
* http://bugzilla.clutter-project.org/show_bug.cgi?id=2621
* https://bugzilla.gnome.org/show_bug.cgi?id=652036
*/
_clutter_actor_traverse (child,
0,
invalidate_queue_redraw_entry,
NULL,
NULL);
}
old_first = self->priv->first_child;
old_last = self->priv->last_child;
remove_child (self, child);
self->priv->n_children -= 1;
self->priv->age += 1;
/* clutter_actor_reparent() will emit ::parent-set for us */
if (emit_parent_set && !CLUTTER_ACTOR_IN_REPARENT (child))
g_signal_emit (child, actor_signals[PARENT_SET], 0, self);
/* if the child was mapped then we need to relayout ourselves to account
* for the removed child
*/
if (was_mapped)
clutter_actor_queue_relayout (self);
/* we need to emit the signal before dropping the reference */
if (emit_actor_removed)
g_signal_emit_by_name (self, "actor-removed", child);
if (notify_first_last)
{
if (old_first != self->priv->first_child)
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_FIRST_CHILD]);
if (old_last != self->priv->last_child)
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_LAST_CHILD]);
}
/* remove the reference we acquired in clutter_actor_add_child() */
g_object_unref (child);
}
static const ClutterTransformInfo default_transform_info = {
0.0, { 0, }, /* rotation-x */
0.0, { 0, }, /* rotation-y */
0.0, { 0, }, /* rotation-z */
1.0, 1.0, { 0, }, /* scale */
{ 0, }, /* anchor */
};
/*< private >
* _clutter_actor_get_transform_info_or_defaults:
* @self: a #ClutterActor
*
* Retrieves the ClutterTransformInfo structure associated to an actor.
*
* If the actor does not have a ClutterTransformInfo structure associated
* to it, then the default structure will be returned.
*
* This function should only be used for getters.
*
* Return value: a const pointer to the ClutterTransformInfo structure
*/
const ClutterTransformInfo *
_clutter_actor_get_transform_info_or_defaults (ClutterActor *self)
{
ClutterTransformInfo *info;
info = g_object_get_qdata (G_OBJECT (self), quark_actor_transform_info);
if (info != NULL)
return info;
return &default_transform_info;
}
static void
clutter_transform_info_free (gpointer data)
{
if (data != NULL)
g_slice_free (ClutterTransformInfo, data);
}
/*< private >
* _clutter_actor_get_transform_info:
* @self: a #ClutterActor
*
* Retrieves a pointer to the ClutterTransformInfo structure.
*
* If the actor does not have a ClutterTransformInfo associated to it, one
* will be created and initialized to the default values.
*
* This function should be used for setters.
*
* For getters, you should use _clutter_actor_get_transform_info_or_defaults()
* instead.
*
* Return value: (transfer none): a pointer to the ClutterTransformInfo
* structure
*/
ClutterTransformInfo *
_clutter_actor_get_transform_info (ClutterActor *self)
{
ClutterTransformInfo *info;
info = g_object_get_qdata (G_OBJECT (self), quark_actor_transform_info);
if (info == NULL)
{
info = g_slice_new (ClutterTransformInfo);
*info = default_transform_info;
g_object_set_qdata_full (G_OBJECT (self), quark_actor_transform_info,
info,
clutter_transform_info_free);
}
return info;
}
/*< private >
* clutter_actor_set_rotation_angle_internal:
* @self: a #ClutterActor
* @axis: the axis of the angle to change
* @angle: the angle of rotation
*
* Sets the rotation angle on the given axis without affecting the
* rotation center point.
*/
static inline void
clutter_actor_set_rotation_angle_internal (ClutterActor *self,
ClutterRotateAxis axis,
gdouble angle)
{
GObject *obj = G_OBJECT (self);
ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info (self);
g_object_freeze_notify (obj);
switch (axis)
{
case CLUTTER_X_AXIS:
info->rx_angle = angle;
g_object_notify_by_pspec (obj, obj_props[PROP_ROTATION_ANGLE_X]);
break;
case CLUTTER_Y_AXIS:
info->ry_angle = angle;
g_object_notify_by_pspec (obj, obj_props[PROP_ROTATION_ANGLE_Y]);
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case CLUTTER_Z_AXIS:
info->rz_angle = angle;
g_object_notify_by_pspec (obj, obj_props[PROP_ROTATION_ANGLE_Z]);
break;
}
self->priv->transform_valid = FALSE;
g_object_thaw_notify (obj);
clutter_actor_queue_redraw (self);
}
/*< private >
* clutter_actor_set_rotation_center_internal:
* @self: a #ClutterActor
* @axis: the axis of the center to change
* @center: the coordinates of the rotation center
*
* Sets the rotation center on the given axis without affecting the
* rotation angle.
*/
static inline void
clutter_actor_set_rotation_center_internal (ClutterActor *self,
ClutterRotateAxis axis,
const ClutterVertex *center)
{
GObject *obj = G_OBJECT (self);
ClutterTransformInfo *info;
ClutterVertex v = { 0, 0, 0 };
info = _clutter_actor_get_transform_info (self);
if (center != NULL)
v = *center;
g_object_freeze_notify (obj);
switch (axis)
{
case CLUTTER_X_AXIS:
clutter_anchor_coord_set_units (&info->rx_center, v.x, v.y, v.z);
g_object_notify_by_pspec (obj, obj_props[PROP_ROTATION_CENTER_X]);
break;
case CLUTTER_Y_AXIS:
clutter_anchor_coord_set_units (&info->ry_center, v.x, v.y, v.z);
g_object_notify_by_pspec (obj, obj_props[PROP_ROTATION_CENTER_Y]);
break;
case CLUTTER_Z_AXIS:
/* if the previously set rotation center was fractional, then
* setting explicit coordinates will have to notify the
* :rotation-center-z-gravity property as well
*/
if (info->rz_center.is_fractional)
g_object_notify_by_pspec (obj, obj_props[PROP_ROTATION_CENTER_Z_GRAVITY]);
clutter_anchor_coord_set_units (&info->rz_center, v.x, v.y, v.z);
g_object_notify_by_pspec (obj, obj_props[PROP_ROTATION_CENTER_Z]);
break;
}
self->priv->transform_valid = FALSE;
g_object_thaw_notify (obj);
clutter_actor_queue_redraw (self);
}
static inline void
clutter_actor_set_scale_factor (ClutterActor *self,
ClutterRotateAxis axis,
gdouble factor)
{
GObject *obj = G_OBJECT (self);
ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info (self);
g_object_freeze_notify (obj);
switch (axis)
{
case CLUTTER_X_AXIS:
info->scale_x = factor;
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_X]);
break;
case CLUTTER_Y_AXIS:
info->scale_y = factor;
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_Y]);
break;
default:
g_assert_not_reached ();
}
self->priv->transform_valid = FALSE;
clutter_actor_queue_redraw (self);
g_object_thaw_notify (obj);
}
static inline void
clutter_actor_set_scale_center (ClutterActor *self,
ClutterRotateAxis axis,
gfloat coord)
{
GObject *obj = G_OBJECT (self);
ClutterTransformInfo *info;
gfloat center_x, center_y;
info = _clutter_actor_get_transform_info (self);
g_object_freeze_notify (obj);
/* get the current scale center coordinates */
clutter_anchor_coord_get_units (self, &info->scale_center,
&center_x,
&center_y,
NULL);
/* we need to notify this too, because setting explicit coordinates will
* change the gravity as a side effect
*/
if (info->scale_center.is_fractional)
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_GRAVITY]);
switch (axis)
{
case CLUTTER_X_AXIS:
clutter_anchor_coord_set_units (&info->scale_center, coord, center_y, 0);
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_CENTER_X]);
break;
case CLUTTER_Y_AXIS:
clutter_anchor_coord_set_units (&info->scale_center, center_x, coord, 0);
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_CENTER_Y]);
break;
default:
g_assert_not_reached ();
}
self->priv->transform_valid = FALSE;
clutter_actor_queue_redraw (self);
g_object_thaw_notify (obj);
}
static inline void
clutter_actor_set_anchor_coord (ClutterActor *self,
ClutterRotateAxis axis,
gfloat coord)
{
GObject *obj = G_OBJECT (self);
ClutterTransformInfo *info;
gfloat anchor_x, anchor_y;
info = _clutter_actor_get_transform_info (self);
g_object_freeze_notify (obj);
clutter_anchor_coord_get_units (self, &info->anchor,
&anchor_x,
&anchor_y,
NULL);
if (info->anchor.is_fractional)
g_object_notify_by_pspec (obj, obj_props[PROP_ANCHOR_GRAVITY]);
switch (axis)
{
case CLUTTER_X_AXIS:
clutter_anchor_coord_set_units (&info->anchor,
coord,
anchor_y,
0.0);
g_object_notify_by_pspec (obj, obj_props[PROP_ANCHOR_X]);
break;
case CLUTTER_Y_AXIS:
clutter_anchor_coord_set_units (&info->anchor,
anchor_x,
coord,
0.0);
g_object_notify_by_pspec (obj, obj_props[PROP_ANCHOR_Y]);
break;
default:
g_assert_not_reached ();
}
self->priv->transform_valid = FALSE;
clutter_actor_queue_redraw (self);
g_object_thaw_notify (obj);
}
static void
clutter_actor_set_property (GObject *object,
guint prop_id,
const GValue *value,
GParamSpec *pspec)
{
ClutterActor *actor = CLUTTER_ACTOR (object);
ClutterActorPrivate *priv = actor->priv;
switch (prop_id)
{
case PROP_X:
clutter_actor_set_x (actor, g_value_get_float (value));
break;
case PROP_Y:
clutter_actor_set_y (actor, g_value_get_float (value));
break;
case PROP_WIDTH:
clutter_actor_set_width (actor, g_value_get_float (value));
break;
case PROP_HEIGHT:
clutter_actor_set_height (actor, g_value_get_float (value));
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_FIXED_X:
clutter_actor_set_x (actor, g_value_get_float (value));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_FIXED_Y:
clutter_actor_set_y (actor, g_value_get_float (value));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_FIXED_POSITION_SET:
clutter_actor_set_fixed_position_set (actor, g_value_get_boolean (value));
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_MIN_WIDTH:
clutter_actor_set_min_width (actor, g_value_get_float (value));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_MIN_HEIGHT:
clutter_actor_set_min_height (actor, g_value_get_float (value));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_NATURAL_WIDTH:
clutter_actor_set_natural_width (actor, g_value_get_float (value));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_NATURAL_HEIGHT:
clutter_actor_set_natural_height (actor, g_value_get_float (value));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_MIN_WIDTH_SET:
clutter_actor_set_min_width_set (actor, g_value_get_boolean (value));
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_MIN_HEIGHT_SET:
clutter_actor_set_min_height_set (actor, g_value_get_boolean (value));
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_NATURAL_WIDTH_SET:
clutter_actor_set_natural_width_set (actor, g_value_get_boolean (value));
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_NATURAL_HEIGHT_SET:
clutter_actor_set_natural_height_set (actor, g_value_get_boolean (value));
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_REQUEST_MODE:
clutter_actor_set_request_mode (actor, g_value_get_enum (value));
break;
case PROP_DEPTH:
clutter_actor_set_depth (actor, g_value_get_float (value));
break;
case PROP_OPACITY:
clutter_actor_set_opacity (actor, g_value_get_uint (value));
break;
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
case PROP_OFFSCREEN_REDIRECT:
clutter_actor_set_offscreen_redirect (actor, g_value_get_enum (value));
break;
case PROP_NAME:
clutter_actor_set_name (actor, g_value_get_string (value));
break;
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
case PROP_VISIBLE:
if (g_value_get_boolean (value) == TRUE)
clutter_actor_show (actor);
else
clutter_actor_hide (actor);
break;
case PROP_SCALE_X:
clutter_actor_set_scale_factor (actor, CLUTTER_X_AXIS,
g_value_get_double (value));
break;
case PROP_SCALE_Y:
clutter_actor_set_scale_factor (actor, CLUTTER_Y_AXIS,
g_value_get_double (value));
break;
case PROP_SCALE_CENTER_X:
clutter_actor_set_scale_center (actor, CLUTTER_X_AXIS,
g_value_get_float (value));
break;
case PROP_SCALE_CENTER_Y:
clutter_actor_set_scale_center (actor, CLUTTER_Y_AXIS,
g_value_get_float (value));
break;
case PROP_SCALE_GRAVITY:
{
const ClutterTransformInfo *info;
ClutterGravity gravity;
info = _clutter_actor_get_transform_info_or_defaults (actor);
gravity = g_value_get_enum (value);
clutter_actor_set_scale_with_gravity (actor,
info->scale_x,
info->scale_y,
gravity);
}
break;
case PROP_CLIP:
{
const ClutterGeometry *geom = g_value_get_boxed (value);
clutter_actor_set_clip (actor,
geom->x, geom->y,
geom->width, geom->height);
}
break;
case PROP_CLIP_TO_ALLOCATION:
clutter_actor_set_clip_to_allocation (actor, g_value_get_boolean (value));
break;
case PROP_REACTIVE:
clutter_actor_set_reactive (actor, g_value_get_boolean (value));
break;
case PROP_ROTATION_ANGLE_X:
clutter_actor_set_rotation_angle_internal (actor,
CLUTTER_X_AXIS,
g_value_get_double (value));
break;
case PROP_ROTATION_ANGLE_Y:
clutter_actor_set_rotation_angle_internal (actor,
CLUTTER_Y_AXIS,
g_value_get_double (value));
break;
case PROP_ROTATION_ANGLE_Z:
clutter_actor_set_rotation_angle_internal (actor,
CLUTTER_Z_AXIS,
g_value_get_double (value));
break;
case PROP_ROTATION_CENTER_X:
clutter_actor_set_rotation_center_internal (actor,
CLUTTER_X_AXIS,
g_value_get_boxed (value));
break;
case PROP_ROTATION_CENTER_Y:
clutter_actor_set_rotation_center_internal (actor,
CLUTTER_Y_AXIS,
g_value_get_boxed (value));
break;
case PROP_ROTATION_CENTER_Z:
clutter_actor_set_rotation_center_internal (actor,
CLUTTER_Z_AXIS,
g_value_get_boxed (value));
break;
case PROP_ROTATION_CENTER_Z_GRAVITY:
{
const ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info_or_defaults (actor);
clutter_actor_set_z_rotation_from_gravity (actor, info->rz_angle,
g_value_get_enum (value));
}
break;
case PROP_ANCHOR_X:
clutter_actor_set_anchor_coord (actor, CLUTTER_X_AXIS,
g_value_get_float (value));
break;
case PROP_ANCHOR_Y:
clutter_actor_set_anchor_coord (actor, CLUTTER_Y_AXIS,
g_value_get_float (value));
break;
case PROP_ANCHOR_GRAVITY:
clutter_actor_set_anchor_point_from_gravity (actor,
g_value_get_enum (value));
break;
case PROP_SHOW_ON_SET_PARENT:
priv->show_on_set_parent = g_value_get_boolean (value);
break;
case PROP_TEXT_DIRECTION:
clutter_actor_set_text_direction (actor, g_value_get_enum (value));
break;
case PROP_ACTIONS:
clutter_actor_add_action (actor, g_value_get_object (value));
break;
case PROP_CONSTRAINTS:
clutter_actor_add_constraint (actor, g_value_get_object (value));
break;
case PROP_EFFECT:
clutter_actor_add_effect (actor, g_value_get_object (value));
break;
case PROP_LAYOUT_MANAGER:
clutter_actor_set_layout_manager (actor, g_value_get_object (value));
break;
case PROP_X_ALIGN:
clutter_actor_set_x_align (actor, g_value_get_enum (value));
break;
case PROP_Y_ALIGN:
clutter_actor_set_y_align (actor, g_value_get_enum (value));
break;
case PROP_MARGIN_TOP:
clutter_actor_set_margin_top (actor, g_value_get_float (value));
break;
case PROP_MARGIN_BOTTOM:
clutter_actor_set_margin_bottom (actor, g_value_get_float (value));
break;
case PROP_MARGIN_LEFT:
clutter_actor_set_margin_left (actor, g_value_get_float (value));
break;
case PROP_MARGIN_RIGHT:
clutter_actor_set_margin_right (actor, g_value_get_float (value));
break;
case PROP_BACKGROUND_COLOR:
clutter_actor_set_background_color (actor, g_value_get_boxed (value));
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
static void
clutter_actor_get_property (GObject *object,
guint prop_id,
GValue *value,
GParamSpec *pspec)
{
ClutterActor *actor = CLUTTER_ACTOR (object);
ClutterActorPrivate *priv = actor->priv;
switch (prop_id)
{
case PROP_X:
g_value_set_float (value, clutter_actor_get_x (actor));
break;
case PROP_Y:
g_value_set_float (value, clutter_actor_get_y (actor));
break;
case PROP_WIDTH:
g_value_set_float (value, clutter_actor_get_width (actor));
break;
case PROP_HEIGHT:
g_value_set_float (value, clutter_actor_get_height (actor));
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_FIXED_X:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->fixed_x);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_FIXED_Y:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->fixed_y);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_FIXED_POSITION_SET:
g_value_set_boolean (value, priv->position_set);
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_MIN_WIDTH:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->min_width);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_MIN_HEIGHT:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->min_height);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_NATURAL_WIDTH:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->natural_width);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_NATURAL_HEIGHT:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->natural_height);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_MIN_WIDTH_SET:
g_value_set_boolean (value, priv->min_width_set);
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_MIN_HEIGHT_SET:
g_value_set_boolean (value, priv->min_height_set);
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_NATURAL_WIDTH_SET:
g_value_set_boolean (value, priv->natural_width_set);
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_NATURAL_HEIGHT_SET:
g_value_set_boolean (value, priv->natural_height_set);
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_REQUEST_MODE:
g_value_set_enum (value, priv->request_mode);
break;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
case PROP_ALLOCATION:
g_value_set_boxed (value, &priv->allocation);
break;
case PROP_DEPTH:
g_value_set_float (value, clutter_actor_get_depth (actor));
break;
case PROP_OPACITY:
g_value_set_uint (value, priv->opacity);
break;
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
case PROP_OFFSCREEN_REDIRECT:
g_value_set_enum (value, priv->offscreen_redirect);
break;
case PROP_NAME:
g_value_set_string (value, priv->name);
break;
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
case PROP_VISIBLE:
g_value_set_boolean (value, CLUTTER_ACTOR_IS_VISIBLE (actor));
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
break;
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
case PROP_MAPPED:
g_value_set_boolean (value, CLUTTER_ACTOR_IS_MAPPED (actor));
break;
case PROP_REALIZED:
g_value_set_boolean (value, CLUTTER_ACTOR_IS_REALIZED (actor));
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
break;
case PROP_HAS_CLIP:
g_value_set_boolean (value, priv->has_clip);
break;
case PROP_CLIP:
{
ClutterGeometry clip;
clip.x = CLUTTER_NEARBYINT (priv->clip.x);
clip.y = CLUTTER_NEARBYINT (priv->clip.y);
clip.width = CLUTTER_NEARBYINT (priv->clip.width);
clip.height = CLUTTER_NEARBYINT (priv->clip.height);
g_value_set_boxed (value, &clip);
}
break;
case PROP_CLIP_TO_ALLOCATION:
g_value_set_boolean (value, priv->clip_to_allocation);
break;
case PROP_SCALE_X:
{
const ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info_or_defaults (actor);
g_value_set_double (value, info->scale_x);
}
break;
case PROP_SCALE_Y:
{
const ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info_or_defaults (actor);
g_value_set_double (value, info->scale_y);
}
break;
case PROP_SCALE_CENTER_X:
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat center;
clutter_actor_get_scale_center (actor, &center, NULL);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_value_set_float (value, center);
}
break;
case PROP_SCALE_CENTER_Y:
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat center;
clutter_actor_get_scale_center (actor, NULL, &center);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_value_set_float (value, center);
}
break;
case PROP_SCALE_GRAVITY:
g_value_set_enum (value, clutter_actor_get_scale_gravity (actor));
break;
case PROP_REACTIVE:
g_value_set_boolean (value, clutter_actor_get_reactive (actor));
break;
case PROP_ROTATION_ANGLE_X:
{
const ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info_or_defaults (actor);
g_value_set_double (value, info->rx_angle);
}
break;
case PROP_ROTATION_ANGLE_Y:
{
const ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info_or_defaults (actor);
g_value_set_double (value, info->ry_angle);
}
break;
case PROP_ROTATION_ANGLE_Z:
{
const ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info_or_defaults (actor);
g_value_set_double (value, info->rz_angle);
}
break;
case PROP_ROTATION_CENTER_X:
{
ClutterVertex center;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_get_rotation (actor, CLUTTER_X_AXIS,
&center.x,
&center.y,
&center.z);
g_value_set_boxed (value, &center);
}
break;
case PROP_ROTATION_CENTER_Y:
{
ClutterVertex center;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_get_rotation (actor, CLUTTER_Y_AXIS,
&center.x,
&center.y,
&center.z);
g_value_set_boxed (value, &center);
}
break;
case PROP_ROTATION_CENTER_Z:
{
ClutterVertex center;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_get_rotation (actor, CLUTTER_Z_AXIS,
&center.x,
&center.y,
&center.z);
g_value_set_boxed (value, &center);
}
break;
case PROP_ROTATION_CENTER_Z_GRAVITY:
g_value_set_enum (value, clutter_actor_get_z_rotation_gravity (actor));
break;
case PROP_ANCHOR_X:
{
const ClutterTransformInfo *info;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat anchor_x;
info = _clutter_actor_get_transform_info_or_defaults (actor);
clutter_anchor_coord_get_units (actor, &info->anchor,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
&anchor_x,
NULL,
NULL);
g_value_set_float (value, anchor_x);
}
break;
case PROP_ANCHOR_Y:
{
const ClutterTransformInfo *info;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat anchor_y;
info = _clutter_actor_get_transform_info_or_defaults (actor);
clutter_anchor_coord_get_units (actor, &info->anchor,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
NULL,
&anchor_y,
NULL);
g_value_set_float (value, anchor_y);
}
break;
case PROP_ANCHOR_GRAVITY:
g_value_set_enum (value, clutter_actor_get_anchor_point_gravity (actor));
break;
case PROP_SHOW_ON_SET_PARENT:
g_value_set_boolean (value, priv->show_on_set_parent);
break;
case PROP_TEXT_DIRECTION:
g_value_set_enum (value, priv->text_direction);
break;
case PROP_HAS_POINTER:
g_value_set_boolean (value, priv->has_pointer);
break;
case PROP_LAYOUT_MANAGER:
g_value_set_object (value, priv->layout_manager);
break;
case PROP_X_ALIGN:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_enum (value, info->x_align);
}
break;
case PROP_Y_ALIGN:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_enum (value, info->y_align);
}
break;
case PROP_MARGIN_TOP:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->margin.top);
}
break;
case PROP_MARGIN_BOTTOM:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->margin.bottom);
}
break;
case PROP_MARGIN_LEFT:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->margin.left);
}
break;
case PROP_MARGIN_RIGHT:
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (actor);
g_value_set_float (value, info->margin.right);
}
break;
case PROP_BACKGROUND_COLOR_SET:
g_value_set_boolean (value, priv->bg_color_set);
break;
case PROP_BACKGROUND_COLOR:
g_value_set_boxed (value, &priv->bg_color);
break;
case PROP_FIRST_CHILD:
g_value_set_object (value, priv->first_child);
break;
case PROP_LAST_CHILD:
g_value_set_object (value, priv->last_child);
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
static void
clutter_actor_dispose (GObject *object)
{
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
ClutterActor *self = CLUTTER_ACTOR (object);
ClutterActorPrivate *priv = self->priv;
CLUTTER_NOTE (MISC, "Disposing of object (id=%d) of type '%s' (ref_count:%d)",
priv->id,
g_type_name (G_OBJECT_TYPE (self)),
object->ref_count);
g_signal_emit (self, actor_signals[DESTROY], 0);
/* avoid recursing when called from clutter_actor_destroy() */
if (priv->parent != NULL)
{
ClutterActor *parent = priv->parent;
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
/* go through the Container implementation unless this
* is an internal child and has been marked as such.
*
* removing the actor from its parent will reset the
* realized and mapped states.
*/
if (!CLUTTER_ACTOR_IS_INTERNAL_CHILD (self))
clutter_container_remove_actor (CLUTTER_CONTAINER (parent), self);
else
clutter_actor_remove_child_internal (parent, self,
REMOVE_CHILD_LEGACY_FLAGS);
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
}
/* parent must be gone at this point */
g_assert (priv->parent == NULL);
if (!CLUTTER_ACTOR_IS_TOPLEVEL (self))
{
/* can't be mapped or realized with no parent */
g_assert (!CLUTTER_ACTOR_IS_MAPPED (self));
g_assert (!CLUTTER_ACTOR_IS_REALIZED (self));
}
g_clear_object (&priv->pango_context);
g_clear_object (&priv->actions);
g_clear_object (&priv->constraints);
g_clear_object (&priv->effects);
g_clear_object (&priv->flatten_effect);
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
2011-11-29 18:10:57 +00:00
if (priv->layout_manager != NULL)
{
clutter_layout_manager_set_container (priv->layout_manager, NULL);
g_object_unref (priv->layout_manager);
priv->layout_manager = NULL;
}
G_OBJECT_CLASS (clutter_actor_parent_class)->dispose (object);
}
static void
clutter_actor_finalize (GObject *object)
{
ClutterActorPrivate *priv = CLUTTER_ACTOR (object)->priv;
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
CLUTTER_NOTE (MISC, "Finalize actor (name='%s', id=%d) of type '%s'",
priv->name != NULL ? priv->name : "<none>",
priv->id,
g_type_name (G_OBJECT_TYPE (object)));
_clutter_context_release_id (priv->id);
g_free (priv->name);
G_OBJECT_CLASS (clutter_actor_parent_class)->finalize (object);
}
/**
* clutter_actor_get_accessible:
* @self: a #ClutterActor
*
* Returns the accessible object that describes the actor to an
* assistive technology.
*
* If no class-specific #AtkObject implementation is available for the
* actor instance in question, it will inherit an #AtkObject
* implementation from the first ancestor class for which such an
* implementation is defined.
*
* The documentation of the <ulink
* url="http://developer.gnome.org/doc/API/2.0/atk/index.html">ATK</ulink>
* library contains more information about accessible objects and
* their uses.
*
* Returns: (transfer none): the #AtkObject associated with @actor
*/
AtkObject *
clutter_actor_get_accessible (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
2010-05-25 12:03:24 +00:00
return CLUTTER_ACTOR_GET_CLASS (self)->get_accessible (self);
}
static AtkObject *
clutter_actor_real_get_accessible (ClutterActor *actor)
{
return atk_gobject_accessible_for_object (G_OBJECT (actor));
}
static AtkObject *
_clutter_actor_ref_accessible (AtkImplementor *implementor)
{
AtkObject *accessible;
accessible = clutter_actor_get_accessible (CLUTTER_ACTOR (implementor));
if (accessible != NULL)
g_object_ref (accessible);
return accessible;
}
static void
atk_implementor_iface_init (AtkImplementorIface *iface)
{
iface->ref_accessible = _clutter_actor_ref_accessible;
}
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
static gboolean
clutter_actor_real_get_paint_volume (ClutterActor *self,
ClutterPaintVolume *volume)
{
ClutterActorPrivate *priv = self->priv;
ClutterActor *child;
gboolean res;
/* this is the default return value: we cannot know if a class
* is going to paint outside its allocation, so we take the
* conservative approach.
*/
res = FALSE;
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
/* we start from the allocation */
clutter_paint_volume_set_from_allocation (volume, self);
/* if the actor has a clip set then we have a pretty definite
* size for the paint volume: the actor cannot possibly paint
* outside the clip region.
*/
if (priv->clip_to_allocation)
{
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
/* the allocation has already been set, so we just flip the
* return value
*/
res = TRUE;
}
else if (priv->has_clip &&
priv->clip.width >= 0 &&
priv->clip.height >= 0)
{
ClutterVertex origin;
origin.x = priv->clip.x;
origin.y = priv->clip.y;
origin.z = 0;
clutter_paint_volume_set_origin (volume, &origin);
clutter_paint_volume_set_width (volume, priv->clip.width);
clutter_paint_volume_set_height (volume, priv->clip.height);
res = TRUE;
}
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
/* if we don't have children we just bail out here... */
if (priv->n_children == 0)
return res;
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
/* ...but if we have children then we ask for their paint volume in
* our coordinates. if any of our children replies that it doesn't
* have a paint volume, we bail out
*/
for (child = priv->first_child;
child != NULL;
child = child->priv->next_sibling)
{
const ClutterPaintVolume *child_volume;
child_volume = clutter_actor_get_transformed_paint_volume (child, self);
if (child_volume == NULL)
{
res = FALSE;
break;
}
clutter_paint_volume_union (volume, child_volume);
res = TRUE;
}
return res;
}
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
static gboolean
clutter_actor_real_has_overlaps (ClutterActor *self)
{
/* By default we'll assume that all actors need an offscreen redirect to get
* the correct opacity. Actors such as ClutterTexture that would never need
* an offscreen redirect can override this to return FALSE. */
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
return TRUE;
}
static void
clutter_actor_real_destroy (ClutterActor *actor)
{
ClutterActorIter iter;
clutter_actor_iter_init (&iter, actor);
while (clutter_actor_iter_next (&iter, NULL))
clutter_actor_iter_destroy (&iter);
}
static GObject *
clutter_actor_constructor (GType gtype,
guint n_props,
GObjectConstructParam *props)
{
GObjectClass *gobject_class;
ClutterActor *self;
GObject *retval;
gobject_class = G_OBJECT_CLASS (clutter_actor_parent_class);
retval = gobject_class->constructor (gtype, n_props, props);
self = CLUTTER_ACTOR (retval);
if (self->priv->layout_manager == NULL)
{
ClutterLayoutManager *default_layout;
CLUTTER_NOTE (LAYOUT, "Creating default layout manager");
default_layout = clutter_fixed_layout_new ();
clutter_actor_set_layout_manager (self, default_layout);
}
return retval;
}
static void
clutter_actor_class_init (ClutterActorClass *klass)
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
quark_shader_data = g_quark_from_static_string ("-clutter-actor-shader-data");
quark_actor_layout_info = g_quark_from_static_string ("-clutter-actor-layout-info");
quark_actor_transform_info = g_quark_from_static_string ("-clutter-actor-transform-info");
object_class->constructor = clutter_actor_constructor;
object_class->set_property = clutter_actor_set_property;
object_class->get_property = clutter_actor_get_property;
object_class->dispose = clutter_actor_dispose;
object_class->finalize = clutter_actor_finalize;
klass->show = clutter_actor_real_show;
klass->show_all = clutter_actor_show;
klass->hide = clutter_actor_real_hide;
klass->hide_all = clutter_actor_hide;
klass->map = clutter_actor_real_map;
klass->unmap = clutter_actor_real_unmap;
klass->unrealize = clutter_actor_real_unrealize;
klass->pick = clutter_actor_real_pick;
klass->get_preferred_width = clutter_actor_real_get_preferred_width;
klass->get_preferred_height = clutter_actor_real_get_preferred_height;
klass->allocate = clutter_actor_real_allocate;
klass->queue_redraw = clutter_actor_real_queue_redraw;
klass->queue_relayout = clutter_actor_real_queue_relayout;
klass->apply_transform = clutter_actor_real_apply_transform;
klass->get_accessible = clutter_actor_real_get_accessible;
klass->get_paint_volume = clutter_actor_real_get_paint_volume;
klass->has_overlaps = clutter_actor_real_has_overlaps;
klass->paint = clutter_actor_real_paint;
klass->destroy = clutter_actor_real_destroy;
g_type_class_add_private (klass, sizeof (ClutterActorPrivate));
/**
* ClutterActor:x:
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* X coordinate of the actor in pixels. If written, forces a fixed
* position for the actor. If read, returns the fixed position if any,
* otherwise the allocation if available, otherwise 0.
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_X] =
g_param_spec_float ("x",
P_("X coordinate"),
P_("X coordinate of the actor"),
-G_MAXFLOAT, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:y:
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Y coordinate of the actor in pixels. If written, forces a fixed
* position for the actor. If read, returns the fixed position if
* any, otherwise the allocation if available, otherwise 0.
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_Y] =
g_param_spec_float ("y",
P_("Y coordinate"),
P_("Y coordinate of the actor"),
-G_MAXFLOAT, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:width:
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Width of the actor (in pixels). If written, forces the minimum and
* natural size request of the actor to the given width. If read, returns
* the allocated width if available, otherwise the width request.
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_WIDTH] =
g_param_spec_float ("width",
P_("Width"),
P_("Width of the actor"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:height:
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Height of the actor (in pixels). If written, forces the minimum and
* natural size request of the actor to the given height. If read, returns
* the allocated height if available, otherwise the height request.
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_HEIGHT] =
g_param_spec_float ("height",
P_("Height"),
P_("Height of the actor"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:fixed-x:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The fixed X position of the actor in pixels.
*
* Writing this property sets #ClutterActor:fixed-position-set
* property as well, as a side effect
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_FIXED_X] =
g_param_spec_float ("fixed-x",
P_("Fixed X"),
P_("Forced X position of the actor"),
-G_MAXFLOAT, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:fixed-y:
*
* The fixed Y position of the actor in pixels.
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Writing this property sets the #ClutterActor:fixed-position-set
* property as well, as a side effect
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_FIXED_Y] =
g_param_spec_float ("fixed-y",
P_("Fixed Y"),
P_("Forced Y position of the actor"),
-G_MAXFLOAT, G_MAXFLOAT,
0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:fixed-position-set:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* This flag controls whether the #ClutterActor:fixed-x and
* #ClutterActor:fixed-y properties are used
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_FIXED_POSITION_SET] =
g_param_spec_boolean ("fixed-position-set",
P_("Fixed position set"),
P_("Whether to use fixed positioning for the actor"),
FALSE,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:min-width:
*
* A forced minimum width request for the actor, in pixels
*
* Writing this property sets the #ClutterActor:min-width-set property
* as well, as a side effect.
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
*This property overrides the usual width request of the actor.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_MIN_WIDTH] =
g_param_spec_float ("min-width",
P_("Min Width"),
P_("Forced minimum width request for the actor"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:min-height:
*
* A forced minimum height request for the actor, in pixels
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Writing this property sets the #ClutterActor:min-height-set property
* as well, as a side effect. This property overrides the usual height
* request of the actor.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_MIN_HEIGHT] =
g_param_spec_float ("min-height",
P_("Min Height"),
P_("Forced minimum height request for the actor"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:natural-width:
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* A forced natural width request for the actor, in pixels
*
* Writing this property sets the #ClutterActor:natural-width-set
* property as well, as a side effect. This property overrides the
* usual width request of the actor
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_NATURAL_WIDTH] =
g_param_spec_float ("natural-width",
P_("Natural Width"),
P_("Forced natural width request for the actor"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:natural-height:
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* A forced natural height request for the actor, in pixels
*
* Writing this property sets the #ClutterActor:natural-height-set
* property as well, as a side effect. This property overrides the
* usual height request of the actor
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_NATURAL_HEIGHT] =
g_param_spec_float ("natural-height",
P_("Natural Height"),
P_("Forced natural height request for the actor"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:min-width-set:
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* This flag controls whether the #ClutterActor:min-width property
* is used
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_MIN_WIDTH_SET] =
g_param_spec_boolean ("min-width-set",
P_("Minimum width set"),
P_("Whether to use the min-width property"),
FALSE,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:min-height-set:
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* This flag controls whether the #ClutterActor:min-height property
* is used
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_MIN_HEIGHT_SET] =
g_param_spec_boolean ("min-height-set",
P_("Minimum height set"),
P_("Whether to use the min-height property"),
FALSE,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:natural-width-set:
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* This flag controls whether the #ClutterActor:natural-width property
* is used
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_NATURAL_WIDTH_SET] =
g_param_spec_boolean ("natural-width-set",
P_("Natural width set"),
P_("Whether to use the natural-width property"),
FALSE,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* ClutterActor:natural-height-set:
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* This flag controls whether the #ClutterActor:natural-height property
* is used
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_NATURAL_HEIGHT_SET] =
g_param_spec_boolean ("natural-height-set",
P_("Natural height set"),
P_("Whether to use the natural-height property"),
FALSE,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* ClutterActor:allocation:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The allocation for the actor, in pixels
*
* This is property is read-only, but you might monitor it to know when an
* actor moves or resizes
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ALLOCATION] =
g_param_spec_boxed ("allocation",
P_("Allocation"),
P_("The actor's allocation"),
CLUTTER_TYPE_ACTOR_BOX,
CLUTTER_PARAM_READABLE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* ClutterActor:request-mode:
*
* Request mode for the #ClutterActor. The request mode determines the
* type of geometry management used by the actor, either height for width
* (the default) or width for height.
*
* For actors implementing height for width, the parent container should get
* the preferred width first, and then the preferred height for that width.
*
* For actors implementing width for height, the parent container should get
* the preferred height first, and then the preferred width for that height.
*
* For instance:
*
* |[
* ClutterRequestMode mode;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* gfloat natural_width, min_width;
* gfloat natural_height, min_height;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* mode = clutter_actor_get_request_mode (child);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* if (mode == CLUTTER_REQUEST_HEIGHT_FOR_WIDTH)
* {
* clutter_actor_get_preferred_width (child, -1,
* &amp;min_width,
* &amp;natural_width);
* clutter_actor_get_preferred_height (child, natural_width,
* &amp;min_height,
* &amp;natural_height);
* }
* else
* {
* clutter_actor_get_preferred_height (child, -1,
* &amp;min_height,
* &amp;natural_height);
* clutter_actor_get_preferred_width (child, natural_height,
* &amp;min_width,
* &amp;natural_width);
* }
* ]|
*
* will retrieve the minimum and natural width and height depending on the
* preferred request mode of the #ClutterActor "child".
*
* The clutter_actor_get_preferred_size() function will implement this
* check for you.
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_REQUEST_MODE] =
g_param_spec_enum ("request-mode",
P_("Request Mode"),
P_("The actor's request mode"),
CLUTTER_TYPE_REQUEST_MODE,
CLUTTER_REQUEST_HEIGHT_FOR_WIDTH,
CLUTTER_PARAM_READWRITE);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* ClutterActor:depth:
*
* The position of the actor on the Z axis
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_DEPTH] =
g_param_spec_float ("depth",
P_("Depth"),
P_("Position on the Z axis"),
-G_MAXFLOAT, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* ClutterActor:opacity:
*
* Opacity of an actor, between 0 (fully transparent) and
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* 255 (fully opaque)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_OPACITY] =
g_param_spec_uint ("opacity",
P_("Opacity"),
P_("Opacity of an actor"),
0, 255,
255,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
/**
* ClutterActor:offscreen-redirect:
*
* Determines the conditions in which the actor will be redirected
* to an offscreen framebuffer while being painted. For example this
* can be used to cache an actor in a framebuffer or for improved
* handling of transparent actors. See
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
* clutter_actor_set_offscreen_redirect() for details.
*
* Since: 1.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_OFFSCREEN_REDIRECT] =
g_param_spec_flags ("offscreen-redirect",
P_("Offscreen redirect"),
P_("Flags controlling when to flatten the actor into a single image"),
CLUTTER_TYPE_OFFSCREEN_REDIRECT,
0,
CLUTTER_PARAM_READWRITE);
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* ClutterActor:visible:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Whether the actor is set to be visible or not
*
* See also #ClutterActor:mapped
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_VISIBLE] =
g_param_spec_boolean ("visible",
P_("Visible"),
P_("Whether the actor is visible or not"),
FALSE,
CLUTTER_PARAM_READWRITE);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/**
* ClutterActor:mapped:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Whether the actor is mapped (will be painted when the stage
* to which it belongs is mapped)
*
* Since: 1.0
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_MAPPED] =
g_param_spec_boolean ("mapped",
P_("Mapped"),
P_("Whether the actor will be painted"),
FALSE,
CLUTTER_PARAM_READABLE);
/**
* ClutterActor:realized:
*
* Whether the actor has been realized
*
* Since: 1.0
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_REALIZED] =
g_param_spec_boolean ("realized",
P_("Realized"),
P_("Whether the actor has been realized"),
FALSE,
CLUTTER_PARAM_READABLE);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* ClutterActor:reactive:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Whether the actor is reactive to events or not
*
* Only reactive actors will emit event-related signals
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_REACTIVE] =
g_param_spec_boolean ("reactive",
P_("Reactive"),
P_("Whether the actor is reactive to events"),
FALSE,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* ClutterActor:has-clip:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Whether the actor has the #ClutterActor:clip property set or not
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_HAS_CLIP] =
g_param_spec_boolean ("has-clip",
P_("Has Clip"),
P_("Whether the actor has a clip set"),
FALSE,
CLUTTER_PARAM_READABLE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:clip:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The clip region for the actor, in actor-relative coordinates
*
* Every part of the actor outside the clip region will not be
* painted
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_CLIP] =
g_param_spec_boxed ("clip",
P_("Clip"),
P_("The clip region for the actor"),
CLUTTER_TYPE_GEOMETRY,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:name:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The name of the actor
*
* Since: 0.2
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_NAME] =
g_param_spec_string ("name",
P_("Name"),
P_("Name of the actor"),
NULL,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:scale-x:
*
* The horizontal scale of the actor
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_SCALE_X] =
g_param_spec_double ("scale-x",
P_("Scale X"),
P_("Scale factor on the X axis"),
0.0, G_MAXDOUBLE,
1.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:scale-y:
*
* The vertical scale of the actor
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_SCALE_Y] =
g_param_spec_double ("scale-y",
P_("Scale Y"),
P_("Scale factor on the Y axis"),
0.0, G_MAXDOUBLE,
1.0,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:scale-center-x:
*
* The horizontal center point for scaling
*
* Since: 1.0
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_SCALE_CENTER_X] =
g_param_spec_float ("scale-center-x",
P_("Scale Center X"),
P_("Horizontal scale center"),
-G_MAXFLOAT, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:scale-center-y:
*
* The vertical center point for scaling
*
* Since: 1.0
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_SCALE_CENTER_Y] =
g_param_spec_float ("scale-center-y",
P_("Scale Center Y"),
P_("Vertical scale center"),
-G_MAXFLOAT, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:scale-gravity:
*
* The center point for scaling expressed as a #ClutterGravity
*
* Since: 1.0
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_SCALE_GRAVITY] =
g_param_spec_enum ("scale-gravity",
P_("Scale Gravity"),
P_("The center of scaling"),
CLUTTER_TYPE_GRAVITY,
CLUTTER_GRAVITY_NONE,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:rotation-angle-x:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The rotation angle on the X axis
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ROTATION_ANGLE_X] =
g_param_spec_double ("rotation-angle-x",
P_("Rotation Angle X"),
P_("The rotation angle on the X axis"),
-G_MAXDOUBLE, G_MAXDOUBLE,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:rotation-angle-y:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The rotation angle on the Y axis
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ROTATION_ANGLE_Y] =
g_param_spec_double ("rotation-angle-y",
P_("Rotation Angle Y"),
P_("The rotation angle on the Y axis"),
-G_MAXDOUBLE, G_MAXDOUBLE,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:rotation-angle-z:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The rotation angle on the Z axis
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ROTATION_ANGLE_Z] =
g_param_spec_double ("rotation-angle-z",
P_("Rotation Angle Z"),
P_("The rotation angle on the Z axis"),
-G_MAXDOUBLE, G_MAXDOUBLE,
0.0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:rotation-center-x:
*
* The rotation center on the X axis.
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ROTATION_CENTER_X] =
g_param_spec_boxed ("rotation-center-x",
P_("Rotation Center X"),
P_("The rotation center on the X axis"),
CLUTTER_TYPE_VERTEX,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:rotation-center-y:
*
* The rotation center on the Y axis.
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ROTATION_CENTER_Y] =
g_param_spec_boxed ("rotation-center-y",
P_("Rotation Center Y"),
P_("The rotation center on the Y axis"),
CLUTTER_TYPE_VERTEX,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:rotation-center-z:
*
* The rotation center on the Z axis.
*
* Since: 0.6
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ROTATION_CENTER_Z] =
g_param_spec_boxed ("rotation-center-z",
P_("Rotation Center Z"),
P_("The rotation center on the Z axis"),
CLUTTER_TYPE_VERTEX,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:rotation-center-z-gravity:
*
* The rotation center on the Z axis expressed as a #ClutterGravity.
*
* Since: 1.0
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ROTATION_CENTER_Z_GRAVITY] =
g_param_spec_enum ("rotation-center-z-gravity",
P_("Rotation Center Z Gravity"),
P_("Center point for rotation around the Z axis"),
CLUTTER_TYPE_GRAVITY,
CLUTTER_GRAVITY_NONE,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:anchor-x:
*
* The X coordinate of an actor's anchor point, relative to
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* the actor coordinate space, in pixels
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ANCHOR_X] =
g_param_spec_float ("anchor-x",
P_("Anchor X"),
P_("X coordinate of the anchor point"),
-G_MAXFLOAT, G_MAXFLOAT,
0,
CLUTTER_PARAM_READWRITE);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/**
* ClutterActor:anchor-y:
*
* The Y coordinate of an actor's anchor point, relative to
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* the actor coordinate space, in pixels
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ANCHOR_Y] =
g_param_spec_float ("anchor-y",
P_("Anchor Y"),
P_("Y coordinate of the anchor point"),
-G_MAXFLOAT, G_MAXFLOAT,
0,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:anchor-gravity:
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The anchor point expressed as a #ClutterGravity
*
* Since: 1.0
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ANCHOR_GRAVITY] =
g_param_spec_enum ("anchor-gravity",
P_("Anchor Gravity"),
P_("The anchor point as a ClutterGravity"),
CLUTTER_TYPE_GRAVITY,
CLUTTER_GRAVITY_NONE,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:show-on-set-parent:
*
* If %TRUE, the actor is automatically shown when parented.
*
* Calling clutter_actor_hide() on an actor which has not been
* parented will set this property to %FALSE as a side effect.
*
* Since: 0.8
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_SHOW_ON_SET_PARENT] =
g_param_spec_boolean ("show-on-set-parent",
P_("Show on set parent"),
P_("Whether the actor is shown when parented"),
TRUE,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:clip-to-allocation:
*
* Whether the clip region should track the allocated area
* of the actor.
*
* This property is ignored if a clip area has been explicitly
* set using clutter_actor_set_clip().
*
* Since: 1.0
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_CLIP_TO_ALLOCATION] =
g_param_spec_boolean ("clip-to-allocation",
P_("Clip to Allocation"),
P_("Sets the clip region to track the actor's allocation"),
FALSE,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:text-direction:
*
* The direction of the text inside a #ClutterActor.
*
* Since: 1.0
*/
obj_props[PROP_TEXT_DIRECTION] =
g_param_spec_enum ("text-direction",
P_("Text Direction"),
P_("Direction of the text"),
CLUTTER_TYPE_TEXT_DIRECTION,
CLUTTER_TEXT_DIRECTION_LTR,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:has-pointer:
*
* Whether the actor contains the pointer of a #ClutterInputDevice
* or not.
*
* Since: 1.2
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_HAS_POINTER] =
g_param_spec_boolean ("has-pointer",
P_("Has Pointer"),
P_("Whether the actor contains the pointer of an input device"),
FALSE,
CLUTTER_PARAM_READABLE);
/**
* ClutterActor:actions:
*
* Adds a #ClutterAction to the actor
*
* Since: 1.4
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_ACTIONS] =
g_param_spec_object ("actions",
P_("Actions"),
P_("Adds an action to the actor"),
CLUTTER_TYPE_ACTION,
CLUTTER_PARAM_WRITABLE);
/**
* ClutterActor:constraints:
*
* Adds a #ClutterConstraint to the actor
*
* Since: 1.4
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_CONSTRAINTS] =
g_param_spec_object ("constraints",
P_("Constraints"),
P_("Adds a constraint to the actor"),
CLUTTER_TYPE_CONSTRAINT,
CLUTTER_PARAM_WRITABLE);
/**
* ClutterActor:effect:
*
* Adds #ClutterEffect to the list of effects be applied on a #ClutterActor
*
* Since: 1.4
*/
2011-11-22 17:03:00 +00:00
obj_props[PROP_EFFECT] =
g_param_spec_object ("effect",
P_("Effect"),
P_("Add an effect to be applied on the actor"),
CLUTTER_TYPE_EFFECT,
CLUTTER_PARAM_WRITABLE);
2011-11-22 17:03:00 +00:00
/**
* ClutterActor:layout-manager:
*
* A delegate object for controlling the layout of the children of
2011-11-22 17:03:00 +00:00
* an actor.
*
* Since: 1.10
*/
obj_props[PROP_LAYOUT_MANAGER] =
g_param_spec_object ("layout-manager",
P_("Layout Manager"),
P_("The object controlling the layout of an actor's children"),
CLUTTER_TYPE_LAYOUT_MANAGER,
CLUTTER_PARAM_READWRITE);
2011-11-22 17:03:00 +00:00
/**
* ClutterActor:x-align:
*
* The alignment of an actor on the X axis, if the actor has been given
* extra space for its allocation.
*
* Since: 1.10
*/
obj_props[PROP_X_ALIGN] =
g_param_spec_enum ("x-align",
P_("X Alignment"),
P_("The alignment of the actor on the X axis within its allocation"),
CLUTTER_TYPE_ACTOR_ALIGN,
CLUTTER_ACTOR_ALIGN_FILL,
CLUTTER_PARAM_READWRITE);
2011-11-22 17:03:00 +00:00
/**
* ClutterActor:y-align:
*
* The alignment of an actor on the Y axis, if the actor has been given
* extra space for its allocation.
*
* Since: 1.10
*/
obj_props[PROP_Y_ALIGN] =
g_param_spec_enum ("y-align",
P_("Y Alignment"),
P_("The alignment of the actor on the Y axis within its allocation"),
CLUTTER_TYPE_ACTOR_ALIGN,
CLUTTER_ACTOR_ALIGN_FILL,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:margin-top:
*
* The margin (in pixels) from the top of the actor.
*
* This property adds a margin to the actor's preferred size; the margin
* will be automatically taken into account when allocating the actor.
*
* Since: 1.10
*/
obj_props[PROP_MARGIN_TOP] =
g_param_spec_float ("margin-top",
P_("Margin Top"),
P_("Extra space at the top"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:margin-bottom:
*
* The margin (in pixels) from the bottom of the actor.
*
* This property adds a margin to the actor's preferred size; the margin
* will be automatically taken into account when allocating the actor.
*
* Since: 1.10
*/
obj_props[PROP_MARGIN_BOTTOM] =
g_param_spec_float ("margin-bottom",
P_("Margin Bottom"),
P_("Extra space at the bottom"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:margin-left:
*
* The margin (in pixels) from the left of the actor.
*
* This property adds a margin to the actor's preferred size; the margin
* will be automatically taken into account when allocating the actor.
*
* Since: 1.10
*/
obj_props[PROP_MARGIN_LEFT] =
g_param_spec_float ("margin-left",
P_("Margin Left"),
P_("Extra space at the left"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:margin-right:
*
* The margin (in pixels) from the right of the actor.
*
* This property adds a margin to the actor's preferred size; the margin
* will be automatically taken into account when allocating the actor.
*
* Since: 1.10
*/
obj_props[PROP_MARGIN_RIGHT] =
g_param_spec_float ("margin-right",
P_("Margin Right"),
P_("Extra space at the right"),
0.0, G_MAXFLOAT,
0.0,
CLUTTER_PARAM_READWRITE);
2011-11-22 17:03:00 +00:00
/**
* ClutterActor:background-color-set:
*
* Whether the #ClutterActor:background-color property has been set.
*
* Since: 1.10
*/
obj_props[PROP_BACKGROUND_COLOR_SET] =
g_param_spec_boolean ("background-color-set",
P_("Background Color Set"),
P_("Whether the background color is set"),
FALSE,
CLUTTER_PARAM_READABLE);
/**
* ClutterActor:background-color:
*
* Paints a solid fill of the actor's allocation using the specified
* color.
*
* Since: 1.10
*/
obj_props[PROP_BACKGROUND_COLOR] =
clutter_param_spec_color ("background-color",
P_("Background color"),
P_("The actor's background color"),
CLUTTER_COLOR_Transparent,
CLUTTER_PARAM_READWRITE);
/**
* ClutterActor:first-child:
*
* The actor's first child.
*
* Since: 1.10
*/
obj_props[PROP_FIRST_CHILD] =
g_param_spec_object ("first-child",
P_("First Child"),
P_("The actor's first child"),
CLUTTER_TYPE_ACTOR,
CLUTTER_PARAM_READABLE);
/**
* ClutterActor:last-child:
*
* The actor's last child.
*
* Since: 1.10
*/
obj_props[PROP_LAST_CHILD] =
g_param_spec_object ("last-child",
P_("Last Child"),
P_("The actor's last child"),
CLUTTER_TYPE_ACTOR,
CLUTTER_PARAM_READABLE);
2011-11-22 17:03:00 +00:00
g_object_class_install_properties (object_class, PROP_LAST, obj_props);
/**
* ClutterActor::destroy:
* @actor: the #ClutterActor which emitted the signal
*
* The ::destroy signal notifies that all references held on the
* actor which emitted it should be released.
*
* The ::destroy signal should be used by all holders of a reference
* on @actor.
*
* This signal might result in the finalization of the #ClutterActor
* if all references are released.
*
* Composite actors and actors implementing the #ClutterContainer
* interface should override the default implementation of the
* class handler of this signal and call clutter_actor_destroy() on
* their children. When overriding the default class handler, it is
* required to chain up to the parent's implementation.
*
* Since: 0.2
*/
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
actor_signals[DESTROY] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("destroy"),
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_CLEANUP | G_SIGNAL_NO_RECURSE | G_SIGNAL_NO_HOOKS,
G_STRUCT_OFFSET (ClutterActorClass, destroy),
NULL, NULL,
_clutter_marshal_VOID__VOID,
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
G_TYPE_NONE, 0);
/**
* ClutterActor::show:
* @actor: the object which received the signal
*
* The ::show signal is emitted when an actor is visible and
* rendered on the stage.
*
* Since: 0.2
*/
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
actor_signals[SHOW] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("show"),
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_FIRST,
G_STRUCT_OFFSET (ClutterActorClass, show),
NULL, NULL,
_clutter_marshal_VOID__VOID,
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
G_TYPE_NONE, 0);
/**
* ClutterActor::hide:
* @actor: the object which received the signal
*
* The ::hide signal is emitted when an actor is no longer rendered
* on the stage.
*
* Since: 0.2
*/
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
actor_signals[HIDE] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("hide"),
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_FIRST,
G_STRUCT_OFFSET (ClutterActorClass, hide),
NULL, NULL,
_clutter_marshal_VOID__VOID,
G_TYPE_NONE, 0);
/**
* ClutterActor::parent-set:
* @actor: the object which received the signal
* @old_parent: (allow-none): the previous parent of the actor, or %NULL
*
* This signal is emitted when the parent of the actor changes.
*
* Since: 0.2
*/
actor_signals[PARENT_SET] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("parent-set"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, parent_set),
NULL, NULL,
_clutter_marshal_VOID__OBJECT,
G_TYPE_NONE, 1,
CLUTTER_TYPE_ACTOR);
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
/**
* ClutterActor::queue-redraw:
* @actor: the actor we're bubbling the redraw request through
* @origin: the actor which initiated the redraw request
*
* The ::queue_redraw signal is emitted when clutter_actor_queue_redraw()
* is called on @origin.
*
* The default implementation for #ClutterActor chains up to the
* parent actor and queues a redraw on the parent, thus "bubbling"
* the redraw queue up through the actor graph. The default
2011-10-17 09:25:07 +00:00
* implementation for #ClutterStage queues a clutter_stage_ensure_redraw()
* in a main loop idle handler.
*
* Note that the @origin actor may be the stage, or a container; it
* does not have to be a leaf node in the actor graph.
*
* Toolkits embedding a #ClutterStage which require a redraw and
* relayout cycle can stop the emission of this signal using the
2011-10-17 09:25:07 +00:00
* GSignal API, redraw the UI and then call clutter_stage_ensure_redraw()
* themselves, like:
*
* |[
* static void
2011-10-17 09:25:07 +00:00
* on_redraw_complete (gpointer data)
* {
2011-10-17 09:25:07 +00:00
* ClutterStage *stage = data;
*
* /&ast; execute the Clutter drawing pipeline &ast;/
2011-10-17 09:25:07 +00:00
* clutter_stage_ensure_redraw (stage);
* }
*
* static void
* on_stage_queue_redraw (ClutterStage *stage)
* {
* /&ast; this prevents the default handler to run &ast;/
* g_signal_stop_emission_by_name (stage, "queue-redraw");
*
* /&ast; queue a redraw with the host toolkit and call
* &ast; a function when the redraw has been completed
* &ast;/
2011-10-17 09:25:07 +00:00
* queue_a_redraw (G_CALLBACK (on_redraw_complete), stage);
* }
* ]|
*
* <note><para>This signal is emitted before the Clutter paint
* pipeline is executed. If you want to know when the pipeline has
* been completed you should connect to the ::paint signal on the
* Stage with g_signal_connect_after().</para></note>
*
* Since: 1.0
*/
actor_signals[QUEUE_REDRAW] =
g_signal_new (I_("queue-redraw"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, queue_redraw),
NULL, NULL,
_clutter_marshal_VOID__OBJECT,
G_TYPE_NONE, 1,
CLUTTER_TYPE_ACTOR);
/**
* ClutterActor::queue-relayout
* @actor: the actor being queued for relayout
*
* The ::queue_layout signal is emitted when clutter_actor_queue_relayout()
* is called on an actor.
*
* The default implementation for #ClutterActor chains up to the
* parent actor and queues a relayout on the parent, thus "bubbling"
* the relayout queue up through the actor graph.
*
* The main purpose of this signal is to allow relayout to be propagated
* properly in the procense of #ClutterClone actors. Applications will
* not normally need to connect to this signal.
*
* Since: 1.2
*/
actor_signals[QUEUE_RELAYOUT] =
g_signal_new (I_("queue-relayout"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, queue_relayout),
NULL, NULL,
_clutter_marshal_VOID__VOID,
G_TYPE_NONE, 0);
/**
* ClutterActor::event:
* @actor: the actor which received the event
* @event: a #ClutterEvent
*
* The ::event signal is emitted each time an event is received
* by the @actor. This signal will be emitted on every actor,
* following the hierarchy chain, until it reaches the top-level
* container (the #ClutterStage).
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::button-press-event:
* @actor: the actor which received the event
* @event: (type ClutterButtonEvent): a #ClutterButtonEvent
*
* The ::button-press-event signal is emitted each time a mouse button
* is pressed on @actor.
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[BUTTON_PRESS_EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("button-press-event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, button_press_event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::button-release-event:
* @actor: the actor which received the event
* @event: (type ClutterButtonEvent): a #ClutterButtonEvent
*
* The ::button-release-event signal is emitted each time a mouse button
* is released on @actor.
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[BUTTON_RELEASE_EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("button-release-event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, button_release_event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::scroll-event:
* @actor: the actor which received the event
* @event: (type ClutterScrollEvent): a #ClutterScrollEvent
*
* The ::scroll-event signal is emitted each time the mouse is
* scrolled on @actor
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[SCROLL_EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("scroll-event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, scroll_event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::key-press-event:
* @actor: the actor which received the event
* @event: (type ClutterKeyEvent): a #ClutterKeyEvent
*
* The ::key-press-event signal is emitted each time a keyboard button
* is pressed while @actor has key focus (see clutter_stage_set_key_focus()).
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[KEY_PRESS_EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("key-press-event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, key_press_event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::key-release-event:
* @actor: the actor which received the event
* @event: (type ClutterKeyEvent): a #ClutterKeyEvent
*
* The ::key-release-event signal is emitted each time a keyboard button
* is released while @actor has key focus (see
* clutter_stage_set_key_focus()).
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[KEY_RELEASE_EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("key-release-event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, key_release_event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::motion-event:
* @actor: the actor which received the event
* @event: (type ClutterMotionEvent): a #ClutterMotionEvent
*
* The ::motion-event signal is emitted each time the mouse pointer is
* moved over @actor.
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[MOTION_EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("motion-event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, motion_event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::key-focus-in:
* @actor: the actor which now has key focus
*
2010-08-17 13:54:20 +00:00
* The ::key-focus-in signal is emitted when @actor receives key focus.
*
* Since: 0.6
*/
actor_signals[KEY_FOCUS_IN] =
g_signal_new (I_("key-focus-in"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, key_focus_in),
NULL, NULL,
_clutter_marshal_VOID__VOID,
G_TYPE_NONE, 0);
/**
* ClutterActor::key-focus-out:
* @actor: the actor which now has key focus
*
* The ::key-focus-out signal is emitted when @actor loses key focus.
*
* Since: 0.6
*/
actor_signals[KEY_FOCUS_OUT] =
g_signal_new (I_("key-focus-out"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, key_focus_out),
NULL, NULL,
_clutter_marshal_VOID__VOID,
G_TYPE_NONE, 0);
/**
* ClutterActor::enter-event:
* @actor: the actor which the pointer has entered.
* @event: (type ClutterCrossingEvent): a #ClutterCrossingEvent
*
* The ::enter-event signal is emitted when the pointer enters the @actor
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[ENTER_EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("enter-event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, enter_event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::leave-event:
* @actor: the actor which the pointer has left
* @event: (type ClutterCrossingEvent): a #ClutterCrossingEvent
*
* The ::leave-event signal is emitted when the pointer leaves the @actor.
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[LEAVE_EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("leave-event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, leave_event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::captured-event:
* @actor: the actor which received the signal
* @event: a #ClutterEvent
*
* The ::captured-event signal is emitted when an event is captured
* by Clutter. This signal will be emitted starting from the top-level
* container (the #ClutterStage) to the actor which received the event
* going down the hierarchy. This signal can be used to intercept every
* event before the specialized events (like
* ClutterActor::button-press-event or ::key-released-event) are
* emitted.
*
* Return value: %TRUE if the event has been handled by the actor,
* or %FALSE to continue the emission.
*
* Since: 0.6
*/
actor_signals[CAPTURED_EVENT] =
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_signal_new (I_("captured-event"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, captured_event),
_clutter_boolean_handled_accumulator, NULL,
_clutter_marshal_BOOLEAN__BOXED,
G_TYPE_BOOLEAN, 1,
CLUTTER_TYPE_EVENT | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::paint:
* @actor: the #ClutterActor that received the signal
*
* The ::paint signal is emitted each time an actor is being painted.
*
* Subclasses of #ClutterActor should override the class signal handler
* and paint themselves in that function.
*
* It is possible to connect a handler to the ::paint signal in order
* to set up some custom aspect of a paint.
*
* Since: 0.8
*/
actor_signals[PAINT] =
g_signal_new (I_("paint"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, paint),
NULL, NULL,
_clutter_marshal_VOID__VOID,
G_TYPE_NONE, 0);
/**
* ClutterActor::realize:
* @actor: the #ClutterActor that received the signal
*
* The ::realize signal is emitted each time an actor is being
* realized.
*
* Since: 0.8
*/
actor_signals[REALIZE] =
g_signal_new (I_("realize"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, realize),
NULL, NULL,
_clutter_marshal_VOID__VOID,
G_TYPE_NONE, 0);
/**
* ClutterActor::unrealize:
* @actor: the #ClutterActor that received the signal
*
* The ::unrealize signal is emitted each time an actor is being
* unrealized.
*
* Since: 0.8
*/
actor_signals[UNREALIZE] =
g_signal_new (I_("unrealize"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, unrealize),
NULL, NULL,
_clutter_marshal_VOID__VOID,
G_TYPE_NONE, 0);
/**
* ClutterActor::pick:
* @actor: the #ClutterActor that received the signal
* @color: the #ClutterColor to be used when picking
*
* The ::pick signal is emitted each time an actor is being painted
* in "pick mode". The pick mode is used to identify the actor during
* the event handling phase, or by clutter_stage_get_actor_at_pos().
* The actor should paint its shape using the passed @pick_color.
*
* Subclasses of #ClutterActor should override the class signal handler
* and paint themselves in that function.
*
* It is possible to connect a handler to the ::pick signal in order
* to set up some custom aspect of a paint in pick mode.
*
* Since: 1.0
*/
actor_signals[PICK] =
g_signal_new (I_("pick"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (ClutterActorClass, pick),
NULL, NULL,
_clutter_marshal_VOID__BOXED,
G_TYPE_NONE, 1,
CLUTTER_TYPE_COLOR | G_SIGNAL_TYPE_STATIC_SCOPE);
/**
* ClutterActor::allocation-changed:
* @actor: the #ClutterActor that emitted the signal
* @box: a #ClutterActorBox with the new allocation
* @flags: #ClutterAllocationFlags for the allocation
*
* The ::allocation-changed signal is emitted when the
* #ClutterActor:allocation property changes. Usually, application
* code should just use the notifications for the :allocation property
* but if you want to track the allocation flags as well, for instance
* to know whether the absolute origin of @actor changed, then you might
* want use this signal instead.
*
* Since: 1.0
*/
actor_signals[ALLOCATION_CHANGED] =
g_signal_new (I_("allocation-changed"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
0,
NULL, NULL,
_clutter_marshal_VOID__BOXED_FLAGS,
G_TYPE_NONE, 2,
CLUTTER_TYPE_ACTOR_BOX,
CLUTTER_TYPE_ALLOCATION_FLAGS);
}
static void
clutter_actor_init (ClutterActor *self)
{
ClutterActorPrivate *priv;
self->priv = priv = CLUTTER_ACTOR_GET_PRIVATE (self);
priv->id = _clutter_context_acquire_id (self);
priv->pick_id = -1;
priv->opacity = 0xff;
priv->show_on_set_parent = TRUE;
priv->needs_width_request = TRUE;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->needs_height_request = TRUE;
priv->needs_allocation = TRUE;
priv->cached_width_age = 1;
priv->cached_height_age = 1;
priv->opacity_override = -1;
priv->enable_model_view_transform = TRUE;
/* Initialize an empty paint volume to start with */
_clutter_paint_volume_init_static (&priv->last_paint_volume, NULL);
priv->last_paint_volume_valid = TRUE;
priv->transform_valid = FALSE;
}
/**
* clutter_actor_new:
*
* Creates a new #ClutterActor.
*
* A newly created actor has a floating reference, which will be sunk
* when it is added to another actor.
*
* Return value: (transfer full): the newly created #ClutterActor
*
* Since: 1.10
*/
ClutterActor *
clutter_actor_new (void)
{
return g_object_new (CLUTTER_TYPE_ACTOR, NULL);
}
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
/**
* clutter_actor_destroy:
* @self: a #ClutterActor
*
* Destroys an actor. When an actor is destroyed, it will break any
* references it holds to other objects. If the actor is inside a
* container, the actor will be removed.
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
*
* When you destroy a container, its children will be destroyed as well.
*
* Note: you cannot destroy the #ClutterStage returned by
* clutter_stage_get_default().
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
*/
void
clutter_actor_destroy (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_object_ref (self);
/* avoid recursion while destroying */
if (!CLUTTER_ACTOR_IN_DESTRUCTION (self))
{
CLUTTER_SET_PRIVATE_FLAGS (self, CLUTTER_IN_DESTRUCTION);
g_object_run_dispose (G_OBJECT (self));
CLUTTER_UNSET_PRIVATE_FLAGS (self, CLUTTER_IN_DESTRUCTION);
}
g_object_unref (self);
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
}
void
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
_clutter_actor_finish_queue_redraw (ClutterActor *self,
ClutterPaintVolume *clip)
{
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
ClutterActorPrivate *priv = self->priv;
ClutterPaintVolume *pv;
gboolean clipped;
2007-08-08 10:20:14 +00:00
/* If we've been explicitly passed a clip volume then there's
* nothing more to calculate, but otherwise the only thing we know
* is that the change is constrained to the given actor.
*
* The idea is that if we know the paint volume for where the actor
* was last drawn (in eye coordinates) and we also have the paint
* volume for where it will be drawn next (in actor coordinates)
* then if we queue a redraw for both these volumes that will cover
* everything that needs to be redrawn to clear the old view and
* show the latest view of the actor.
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
*
* Don't clip this redraw if we don't know what position we had for
* the previous redraw since we don't know where to set the clip so
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
* it will clear the actor as it is currently.
*/
if (clip)
{
_clutter_actor_set_queue_redraw_clip (self, clip);
clipped = TRUE;
}
else if (G_LIKELY (priv->last_paint_volume_valid))
{
pv = _clutter_actor_get_paint_volume_mutable (self);
if (pv)
{
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
ClutterActor *stage = _clutter_actor_get_stage_internal (self);
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
/* make sure we redraw the actors old position... */
_clutter_actor_set_queue_redraw_clip (stage,
&priv->last_paint_volume);
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
_clutter_actor_signal_queue_redraw (stage, stage);
_clutter_actor_set_queue_redraw_clip (stage, NULL);
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
/* XXX: Ideally the redraw signal would take a clip volume
* argument, but that would be an ABI break. Until we can
* break the ABI we pass the argument out-of-band
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
*/
/* setup the clip for the actors new position... */
_clutter_actor_set_queue_redraw_clip (self, pv);
clipped = TRUE;
}
else
clipped = FALSE;
}
else
clipped = FALSE;
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
_clutter_actor_signal_queue_redraw (self, self);
/* Just in case anyone is manually firing redraw signals without
* using the public queue_redraw() API we are careful to ensure that
* our out-of-band clip member is cleared before returning...
*
* Note: A NULL clip denotes a full-stage, un-clipped redraw
*/
if (G_LIKELY (clipped))
_clutter_actor_set_queue_redraw_clip (self, NULL);
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
priv->queue_redraw_entry = NULL;
}
static void
_clutter_actor_get_allocation_clip (ClutterActor *self,
ClutterActorBox *clip)
{
ClutterActorBox allocation;
/* XXX: we don't care if we get an out of date allocation here
* because clutter_actor_queue_redraw_with_clip knows to ignore
* the clip if the actor's allocation is invalid.
*
* This is noted because clutter_actor_get_allocation_box does some
* unnecessary work to support buggy code with a comment suggesting
* that it could be changed later which would be good for this use
* case!
*/
clutter_actor_get_allocation_box (self, &allocation);
/* NB: clutter_actor_queue_redraw_with_clip expects a box in the
* actor's own coordinate space but the allocation is in parent
* coordinates */
clip->x1 = 0;
clip->y1 = 0;
clip->x2 = allocation.x2 - allocation.x1;
clip->y2 = allocation.y2 - allocation.y1;
}
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
void
_clutter_actor_queue_redraw_full (ClutterActor *self,
ClutterRedrawFlags flags,
ClutterPaintVolume *volume,
ClutterEffect *effect)
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
{
ClutterActorPrivate *priv = self->priv;
ClutterPaintVolume allocation_pv;
ClutterPaintVolume *pv;
gboolean should_free_pv;
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
ClutterActor *stage;
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
/* Here's an outline of the actor queue redraw mechanism:
*
* The process starts in one of the following two functions which
* are wrappers for this function:
* clutter_actor_queue_redraw
* _clutter_actor_queue_redraw_with_clip
*
* additionally, an effect can queue a redraw by wrapping this
* function in clutter_effect_queue_rerun
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
*
* This functions queues an entry in a list associated with the
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
* stage which is a list of actors that queued a redraw while
* updating the timelines, performing layouting and processing other
* mainloop sources before the next paint starts.
*
* We aim to minimize the processing done at this point because
* there is a good chance other events will happen while updating
* the scenegraph that would invalidate any expensive work we might
* otherwise try to do here. For example we don't try and resolve
* the screen space bounding box of an actor at this stage so as to
* minimize how much of the screen redraw because it's possible
* something else will happen which will force a full redraw anyway.
*
* When all updates are complete and we come to paint the stage then
* we iterate this list and actually emit the "queue-redraw" signals
* for each of the listed actors which will bubble up to the stage
* for each actor and at that point we will transform the actors
* paint volume into screen coordinates to determine the clip region
* for what needs to be redrawn in the next paint.
*
* Besides minimizing redundant work another reason for this
* deferred design is that it's more likely we will be able to
* determine the paint volume of an actor once we've finished
* updating the scenegraph because its allocation should be up to
* date. NB: If we can't determine an actors paint volume then we
* can't automatically queue a clipped redraw which can make a big
* difference to performance.
*
* So the control flow goes like this:
* One of clutter_actor_queue_redraw,
* _clutter_actor_queue_redraw_with_clip
* or clutter_effect_queue_rerun
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
*
* then control moves to:
* _clutter_stage_queue_actor_redraw
*
* later during _clutter_stage_do_update, once relayouting is done
* and the scenegraph has been updated we will call:
* _clutter_stage_finish_queue_redraws
*
* _clutter_stage_finish_queue_redraws will call
* _clutter_actor_finish_queue_redraw for each listed actor.
* Note: actors *are* allowed to queue further redraws during this
* process (considering clone actors or texture_new_from_actor which
* respond to their source queueing a redraw by queuing a redraw
* themselves). We repeat the process until the list is empty.
*
* This will result in the "queue-redraw" signal being fired for
* each actor which will pass control to the default signal handler:
* clutter_actor_real_queue_redraw
*
* This will bubble up to the stages handler:
* clutter_stage_real_queue_redraw
*
* clutter_stage_real_queue_redraw will transform the actors paint
* volume into screen space and add it as a clip region for the next
* paint.
*/
/* ignore queueing a redraw for actors being destroyed */
if (CLUTTER_ACTOR_IN_DESTRUCTION (self))
return;
stage = _clutter_actor_get_stage_internal (self);
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
/* Ignore queueing a redraw for actors not descended from a stage */
if (stage == NULL)
actor: defer queue-redraw signaling Instead of immediately, recursively emitting the "queue-redraw" signal when clutter_actor_queue_redraw is called we now defer this process until all stage updates are complete. This allows us to aggregate repeated _queue_redraw requests for the same actor avoiding redundant paint volume transformations. By deferring we also increase the likelihood that the actor will have a valid paint volume since it will have an up to date allocation; this in turn means we will more often be able to automatically queue clipped redraws which can have a big impact on performance. Here's an outline of the actor queue redraw mechanism: The process starts in clutter_actor_queue_redraw or _clutter_actor_queue_redraw_with_clip. These functions queue an entry in a list associated with the stage which is a list of actors that queued a redraw while updating the timelines, performing layouting and processing other mainloop sources before the next paint starts. We aim to minimize the processing done at this point because there is a good chance other events will happen while updating the scenegraph that would invalidate any expensive work we might otherwise try to do here. For example we don't try and resolve the screen space bounding box of an actor at this stage so as to minimize how much of the screen redraw because it's possible something else will happen which will force a full redraw anyway. When all updates are complete and we come to paint the stage (see _clutter_stage_do_update) then we iterate this list and actually emit the "queue-redraw" signals for each of the listed actors which will bubble up to the stage for each actor and at that point we will transform the actors paint volume into screen coordinates to determine the clip region for what needs to be redrawn in the next paint. Note: actors are allowed to queue a redraw in reseponse to a queue-redraw signal so we repeat the processing of the list until it remains empty. An example of when this happens is for Clone actors or clutter_texture_new_from_actor actors which need to queue a redraw if their source queues a redraw.
2010-09-10 00:33:02 +00:00
return;
/* ignore queueing a redraw on stages that are being destroyed */
if (CLUTTER_ACTOR_IN_DESTRUCTION (stage))
return;
if (flags & CLUTTER_REDRAW_CLIPPED_TO_ALLOCATION)
{
ClutterActorBox allocation_clip;
ClutterVertex origin;
/* If the actor doesn't have a valid allocation then we will
* queue a full stage redraw. */
if (priv->needs_allocation)
{
/* NB: NULL denotes an undefined clip which will result in a
* full redraw... */
_clutter_actor_set_queue_redraw_clip (self, NULL);
_clutter_actor_signal_queue_redraw (self, self);
return;
}
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
_clutter_paint_volume_init_static (&allocation_pv, self);
pv = &allocation_pv;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
_clutter_actor_get_allocation_clip (self, &allocation_clip);
origin.x = allocation_clip.x1;
origin.y = allocation_clip.y1;
origin.z = 0;
clutter_paint_volume_set_origin (pv, &origin);
clutter_paint_volume_set_width (pv,
allocation_clip.x2 - allocation_clip.x1);
clutter_paint_volume_set_height (pv,
allocation_clip.y2 -
allocation_clip.y1);
should_free_pv = TRUE;
}
else
{
pv = volume;
should_free_pv = FALSE;
}
self->priv->queue_redraw_entry =
_clutter_stage_queue_actor_redraw (CLUTTER_STAGE (stage),
priv->queue_redraw_entry,
self,
pv);
if (should_free_pv)
clutter_paint_volume_free (pv);
/* If this is the first redraw queued then we can directly use the
effect parameter */
if (!priv->is_dirty)
priv->effect_to_redraw = effect;
/* Otherwise we need to merge it with the existing effect parameter */
else if (effect != NULL)
{
/* If there's already an effect then we need to use whichever is
later in the chain of actors. Otherwise a full redraw has
already been queued on the actor so we need to ignore the
effect parameter */
if (priv->effect_to_redraw != NULL)
{
if (priv->effects == NULL)
g_warning ("Redraw queued with an effect that is "
"not applied to the actor");
else
{
const GList *l;
for (l = _clutter_meta_group_peek_metas (priv->effects);
l != NULL;
l = l->next)
{
if (l->data == priv->effect_to_redraw ||
l->data == effect)
priv->effect_to_redraw = l->data;
}
}
}
}
else
{
/* If no effect is specified then we need to redraw the whole
actor */
priv->effect_to_redraw = NULL;
}
priv->is_dirty = TRUE;
}
/**
* clutter_actor_queue_redraw:
* @self: A #ClutterActor
*
* Queues up a redraw of an actor and any children. The redraw occurs
* once the main loop becomes idle (after the current batch of events
* has been processed, roughly).
*
* Applications rarely need to call this, as redraws are handled
* automatically by modification functions.
*
* This function will not do anything if @self is not visible, or
* if the actor is inside an invisible part of the scenegraph.
*
* Also be aware that painting is a NOP for actors with an opacity of
* 0
*
* When you are implementing a custom actor you must queue a redraw
* whenever some private state changes that will affect painting or
* picking of your actor.
*/
void
clutter_actor_queue_redraw (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
_clutter_actor_queue_redraw_full (self,
0, /* flags */
NULL, /* clip volume */
NULL /* effect */);
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
}
/*< private >
* _clutter_actor_queue_redraw_with_clip:
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
* @self: A #ClutterActor
* @flags: A mask of #ClutterRedrawFlags controlling the behaviour of
* this queue redraw.
* @volume: A #ClutterPaintVolume describing the bounds of what needs to be
* redrawn or %NULL if you are just using a @flag to state your
* desired clipping.
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
*
* Queues up a clipped redraw of an actor and any children. The redraw
* occurs once the main loop becomes idle (after the current batch of
* events has been processed, roughly).
*
* If no flags are given the clip volume is defined by @volume
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
* specified in actor coordinates and tells Clutter that only content
* within this volume has been changed so Clutter can optionally
* optimize the redraw.
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
*
* If the %CLUTTER_REDRAW_CLIPPED_TO_ALLOCATION @flag is used, @volume
* should be %NULL and this tells Clutter to use the actor's current
* allocation as a clip box. This flag can only be used for 2D actors,
* because any actor with depth may be projected outside its
* allocation.
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
*
* Applications rarely need to call this, as redraws are handled
* automatically by modification functions.
*
* This function will not do anything if @self is not visible, or if
* the actor is inside an invisible part of the scenegraph.
*
* Also be aware that painting is a NOP for actors with an opacity of
* 0
*
* When you are implementing a custom actor you must queue a redraw
* whenever some private state changes that will affect painting or
* picking of your actor.
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
*/
void
_clutter_actor_queue_redraw_with_clip (ClutterActor *self,
ClutterRedrawFlags flags,
ClutterPaintVolume *volume)
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
{
_clutter_actor_queue_redraw_full (self,
flags, /* flags */
volume, /* clip volume */
NULL /* effect */);
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
}
static void
_clutter_actor_queue_only_relayout (ClutterActor *self)
{
ClutterActorPrivate *priv = self->priv;
if (CLUTTER_ACTOR_IN_DESTRUCTION (self))
return;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (priv->needs_width_request &&
priv->needs_height_request &&
priv->needs_allocation)
return; /* save some cpu cycles */
#if CLUTTER_ENABLE_DEBUG
if (!CLUTTER_ACTOR_IS_TOPLEVEL (self) && CLUTTER_ACTOR_IN_RELAYOUT (self))
{
g_warning ("The actor '%s' is currently inside an allocation "
"cycle; calling clutter_actor_queue_relayout() is "
"not recommended",
_clutter_actor_get_debug_name (self));
}
#endif /* CLUTTER_ENABLE_DEBUG */
g_signal_emit (self, actor_signals[QUEUE_RELAYOUT], 0);
}
/**
* clutter_actor_queue_redraw_with_clip:
* @self: a #ClutterActor
* @clip: (allow-none): a rectangular clip region, or %NULL
*
* Queues a redraw on @self limited to a specific, actor-relative
* rectangular area.
*
* If @clip is %NULL this function is equivalent to
* clutter_actor_queue_redraw().
*
* Since: 1.10
*/
void
clutter_actor_queue_redraw_with_clip (ClutterActor *self,
const cairo_rectangle_int_t *clip)
{
ClutterPaintVolume volume;
ClutterVertex origin;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (clip == NULL)
{
clutter_actor_queue_redraw (self);
return;
}
_clutter_paint_volume_init_static (&volume, self);
origin.x = clip->x;
origin.y = clip->y;
origin.z = 0.0f;
clutter_paint_volume_set_origin (&volume, &origin);
clutter_paint_volume_set_width (&volume, clip->width);
clutter_paint_volume_set_height (&volume, clip->height);
_clutter_actor_queue_redraw_full (self, 0, &volume, NULL);
clutter_paint_volume_free (&volume);
}
/**
* clutter_actor_queue_relayout:
* @self: A #ClutterActor
*
* Indicates that the actor's size request or other layout-affecting
* properties may have changed. This function is used inside #ClutterActor
* subclass implementations, not by applications directly.
*
* Queueing a new layout automatically queues a redraw as well.
*
* Since: 0.8
*/
void
clutter_actor_queue_relayout (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
_clutter_actor_queue_only_relayout (self);
clutter_actor_queue_redraw (self);
}
/**
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* clutter_actor_get_preferred_size:
* @self: a #ClutterActor
2009-07-02 13:21:16 +00:00
* @min_width_p: (out) (allow-none): return location for the minimum
* width, or %NULL
* @min_height_p: (out) (allow-none): return location for the minimum
* height, or %NULL
* @natural_width_p: (out) (allow-none): return location for the natural
* width, or %NULL
* @natural_height_p: (out) (allow-none): return location for the natural
* height, or %NULL
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Computes the preferred minimum and natural size of an actor, taking into
* account the actor's geometry management (either height-for-width
* or width-for-height).
*
* The width and height used to compute the preferred height and preferred
* width are the actor's natural ones.
*
* If you need to control the height for the preferred width, or the width for
* the preferred height, you should use clutter_actor_get_preferred_width()
* and clutter_actor_get_preferred_height(), and check the actor's preferred
* geometry management using the #ClutterActor:request-mode property.
*
* Since: 0.8
*/
void
clutter_actor_get_preferred_size (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat *min_width_p,
gfloat *min_height_p,
gfloat *natural_width_p,
gfloat *natural_height_p)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
ClutterActorPrivate *priv;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat min_width, min_height;
gfloat natural_width, natural_height;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
min_width = min_height = 0;
natural_width = natural_height = 0;
if (priv->request_mode == CLUTTER_REQUEST_HEIGHT_FOR_WIDTH)
{
CLUTTER_NOTE (LAYOUT, "Preferred size (height-for-width)");
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_preferred_width (self, -1,
&min_width,
&natural_width);
clutter_actor_get_preferred_height (self, natural_width,
&min_height,
&natural_height);
}
else
{
CLUTTER_NOTE (LAYOUT, "Preferred size (width-for-height)");
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_preferred_height (self, -1,
&min_height,
&natural_height);
clutter_actor_get_preferred_width (self, natural_height,
&min_width,
&natural_width);
}
if (min_width_p)
*min_width_p = min_width;
if (min_height_p)
*min_height_p = min_height;
if (natural_width_p)
*natural_width_p = natural_width;
if (natural_height_p)
*natural_height_p = natural_height;
}
/*< private >
* effective_align:
* @align: a #ClutterActorAlign
* @direction: a #ClutterTextDirection
*
* Retrieves the correct alignment depending on the text direction
*
* Return value: the effective alignment
*/
static ClutterActorAlign
effective_align (ClutterActorAlign align,
ClutterTextDirection direction)
{
ClutterActorAlign res;
switch (align)
{
case CLUTTER_ACTOR_ALIGN_START:
res = (direction == CLUTTER_TEXT_DIRECTION_RTL)
? CLUTTER_ACTOR_ALIGN_END
: CLUTTER_ACTOR_ALIGN_START;
break;
case CLUTTER_ACTOR_ALIGN_END:
res = (direction == CLUTTER_TEXT_DIRECTION_RTL)
? CLUTTER_ACTOR_ALIGN_START
: CLUTTER_ACTOR_ALIGN_END;
break;
default:
res = align;
break;
}
return res;
}
static inline void
adjust_for_margin (float margin_start,
float margin_end,
float *minimum_size,
float *natural_size,
float *allocated_start,
float *allocated_end)
{
*minimum_size -= (margin_start + margin_end);
*natural_size -= (margin_start + margin_end);
*allocated_start += margin_start;
*allocated_end -= margin_end;
}
static inline void
adjust_for_alignment (ClutterActorAlign alignment,
float natural_size,
float *allocated_start,
float *allocated_end)
{
float allocated_size = *allocated_end - *allocated_start;
switch (alignment)
{
case CLUTTER_ACTOR_ALIGN_FILL:
/* do nothing */
break;
case CLUTTER_ACTOR_ALIGN_START:
/* keep start */
*allocated_end = *allocated_start + MIN (natural_size, allocated_size);
break;
case CLUTTER_ACTOR_ALIGN_END:
if (allocated_size > natural_size)
{
*allocated_start += (allocated_size - natural_size);
*allocated_end = *allocated_start + natural_size;
}
break;
case CLUTTER_ACTOR_ALIGN_CENTER:
if (allocated_size > natural_size)
{
*allocated_start += ceilf ((allocated_size - natural_size) / 2);
*allocated_end = *allocated_start + MIN (allocated_size, natural_size);
}
break;
}
}
/*< private >
* clutter_actor_adjust_width:
* @self: a #ClutterActor
* @minimum_width: (inout): the actor's preferred minimum width, which
* will be adjusted depending on the margin
* @natural_width: (inout): the actor's preferred natural width, which
* will be adjusted depending on the margin
* @adjusted_x1: (out): the adjusted x1 for the actor's bounding box
* @adjusted_x2: (out): the adjusted x2 for the actor's bounding box
*
* Adjusts the preferred and allocated position and size of an actor,
* depending on the margin and alignment properties.
*/
static void
clutter_actor_adjust_width (ClutterActor *self,
gfloat *minimum_width,
gfloat *natural_width,
gfloat *adjusted_x1,
gfloat *adjusted_x2)
{
ClutterTextDirection text_dir;
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (self);
text_dir = clutter_actor_get_text_direction (self);
CLUTTER_NOTE (LAYOUT, "Adjusting allocated X and width");
/* this will tweak natural_width to remove the margin, so that
* adjust_for_alignment() will use the correct size
*/
adjust_for_margin (info->margin.left, info->margin.right,
minimum_width, natural_width,
adjusted_x1, adjusted_x2);
adjust_for_alignment (effective_align (info->x_align, text_dir),
*natural_width,
adjusted_x1, adjusted_x2);
}
/*< private >
* clutter_actor_adjust_height:
* @self: a #ClutterActor
* @minimum_height: (inout): the actor's preferred minimum height, which
* will be adjusted depending on the margin
* @natural_height: (inout): the actor's preferred natural height, which
* will be adjusted depending on the margin
* @adjusted_y1: (out): the adjusted y1 for the actor's bounding box
* @adjusted_y2: (out): the adjusted y2 for the actor's bounding box
*
* Adjusts the preferred and allocated position and size of an actor,
* depending on the margin and alignment properties.
*/
static void
clutter_actor_adjust_height (ClutterActor *self,
gfloat *minimum_height,
gfloat *natural_height,
gfloat *adjusted_y1,
gfloat *adjusted_y2)
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (self);
CLUTTER_NOTE (LAYOUT, "Adjusting allocated Y and height");
/* this will tweak natural_height to remove the margin, so that
* adjust_for_alignment() will use the correct size
*/
adjust_for_margin (info->margin.top, info->margin.bottom,
minimum_height, natural_height,
adjusted_y1,
adjusted_y2);
/* we don't use effective_align() here, because text direction
* only affects the horizontal axis
*/
adjust_for_alignment (info->y_align,
*natural_height,
adjusted_y1,
adjusted_y2);
}
/* looks for a cached size request for this for_size. If not
* found, returns the oldest entry so it can be overwritten */
static gboolean
_clutter_actor_get_cached_size_request (gfloat for_size,
SizeRequest *cached_size_requests,
SizeRequest **result)
{
guint i;
*result = &cached_size_requests[0];
for (i = 0; i < N_CACHED_SIZE_REQUESTS; i++)
{
SizeRequest *sr;
sr = &cached_size_requests[i];
if (sr->age > 0 &&
sr->for_size == for_size)
{
CLUTTER_NOTE (LAYOUT, "Size cache hit for size: %.2f", for_size);
*result = sr;
return TRUE;
}
else if (sr->age < (*result)->age)
{
*result = sr;
}
}
CLUTTER_NOTE (LAYOUT, "Size cache miss for size: %.2f", for_size);
return FALSE;
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* clutter_actor_get_preferred_width:
* @self: A #ClutterActor
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* @for_height: available height when computing the preferred width,
* or a negative value to indicate that no height is defined
2009-07-02 13:21:16 +00:00
* @min_width_p: (out) (allow-none): return location for minimum width,
* or %NULL
* @natural_width_p: (out) (allow-none): return location for the natural
* width, or %NULL
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Computes the requested minimum and natural widths for an actor,
* optionally depending on the specified height, or if they are
* already computed, returns the cached values.
*
* An actor may not get its request - depending on the layout
* manager that's in effect.
*
* A request should not incorporate the actor's scale or anchor point;
* those transformations do not affect layout, only rendering.
*
* Since: 0.8
*/
void
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_preferred_width (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat for_height,
gfloat *min_width_p,
gfloat *natural_width_p)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
float request_min_width, request_natural_width;
SizeRequest *cached_size_request;
const ClutterLayoutInfo *info;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
ClutterActorPrivate *priv;
gboolean found_in_cache;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
info = _clutter_actor_get_layout_info_or_defaults (self);
/* we shortcircuit the case of a fixed size set using set_width() */
if (priv->min_width_set && priv->natural_width_set)
{
if (min_width_p != NULL)
*min_width_p = info->min_width + (info->margin.left + info->margin.right);
if (natural_width_p != NULL)
*natural_width_p = info->natural_width + (info->margin.left + info->margin.right);
return;
}
/* the remaining cases are:
*
* - either min_width or natural_width have been set
* - neither min_width or natural_width have been set
*
* in both cases, we go through the cache (and through the actor in case
* of cache misses) and determine the authoritative value depending on
* the *_set flags.
*/
if (!priv->needs_width_request)
{
found_in_cache =
_clutter_actor_get_cached_size_request (for_height,
priv->width_requests,
&cached_size_request);
}
else
{
/* if the actor needs a width request we use the first slot */
found_in_cache = FALSE;
cached_size_request = &priv->width_requests[0];
}
if (!found_in_cache)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
gfloat minimum_width, natural_width;
ClutterActorClass *klass;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
minimum_width = natural_width = 0;
/* adjust for the margin */
if (for_height >= 0)
{
for_height -= (info->margin.top + info->margin.bottom);
if (for_height < 0)
for_height = 0;
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
CLUTTER_NOTE (LAYOUT, "Width request for %.2f px", for_height);
klass = CLUTTER_ACTOR_GET_CLASS (self);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
klass->get_preferred_width (self, for_height,
&minimum_width,
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
&natural_width);
/* adjust for the margin */
minimum_width += (info->margin.left + info->margin.right);
natural_width += (info->margin.left + info->margin.right);
/* Due to accumulated float errors, it's better not to warn
* on this, but just fix it.
*/
if (natural_width < minimum_width)
natural_width = minimum_width;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
cached_size_request->min_size = minimum_width;
cached_size_request->natural_size = natural_width;
cached_size_request->for_size = for_height;
cached_size_request->age = priv->cached_width_age;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->cached_width_age += 1;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->needs_width_request = FALSE;
}
if (!priv->min_width_set)
request_min_width = cached_size_request->min_size;
else
request_min_width = info->min_width;
if (!priv->natural_width_set)
request_natural_width = cached_size_request->natural_size;
else
request_natural_width = info->natural_width;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (min_width_p)
*min_width_p = request_min_width;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (natural_width_p)
*natural_width_p = request_natural_width;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
}
/**
* clutter_actor_get_preferred_height:
* @self: A #ClutterActor
* @for_width: available width to assume in computing desired height,
* or a negative value to indicate that no width is defined
2009-07-02 13:21:16 +00:00
* @min_height_p: (out) (allow-none): return location for minimum height,
* or %NULL
* @natural_height_p: (out) (allow-none): return location for natural
* height, or %NULL
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Computes the requested minimum and natural heights for an actor,
* or if they are already computed, returns the cached values.
*
* An actor may not get its request - depending on the layout
* manager that's in effect.
*
* A request should not incorporate the actor's scale or anchor point;
* those transformations do not affect layout, only rendering.
*
* Since: 0.8
*/
void
clutter_actor_get_preferred_height (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat for_width,
gfloat *min_height_p,
gfloat *natural_height_p)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
float request_min_height, request_natural_height;
SizeRequest *cached_size_request;
const ClutterLayoutInfo *info;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
ClutterActorPrivate *priv;
gboolean found_in_cache;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
info = _clutter_actor_get_layout_info_or_defaults (self);
/* we shortcircuit the case of a fixed size set using set_height() */
if (priv->min_height_set && priv->natural_height_set)
{
if (min_height_p != NULL)
*min_height_p = info->min_height + (info->margin.top + info->margin.bottom);
if (natural_height_p != NULL)
*natural_height_p = info->natural_height + (info->margin.top + info->margin.bottom);
return;
}
/* the remaining cases are:
*
* - either min_height or natural_height have been set
* - neither min_height or natural_height have been set
*
* in both cases, we go through the cache (and through the actor in case
* of cache misses) and determine the authoritative value depending on
* the *_set flags.
*/
if (!priv->needs_height_request)
{
found_in_cache =
_clutter_actor_get_cached_size_request (for_width,
priv->height_requests,
&cached_size_request);
}
else
{
found_in_cache = FALSE;
cached_size_request = &priv->height_requests[0];
}
if (!found_in_cache)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
gfloat minimum_height, natural_height;
ClutterActorClass *klass;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
minimum_height = natural_height = 0;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
CLUTTER_NOTE (LAYOUT, "Height request for %.2f px", for_width);
/* adjust for margin */
if (for_width >= 0)
{
for_width -= (info->margin.left + info->margin.right);
if (for_width < 0)
for_width = 0;
}
klass = CLUTTER_ACTOR_GET_CLASS (self);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
klass->get_preferred_height (self, for_width,
&minimum_height,
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
&natural_height);
/* adjust for margin */
minimum_height += (info->margin.top + info->margin.bottom);
natural_height += (info->margin.top + info->margin.bottom);
/* Due to accumulated float errors, it's better not to warn
* on this, but just fix it.
*/
if (natural_height < minimum_height)
natural_height = minimum_height;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
cached_size_request->min_size = minimum_height;
cached_size_request->natural_size = natural_height;
cached_size_request->for_size = for_width;
cached_size_request->age = priv->cached_height_age;
priv->cached_height_age += 1;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->needs_height_request = FALSE;
}
if (!priv->min_height_set)
request_min_height = cached_size_request->min_size;
else
request_min_height = info->min_height;
if (!priv->natural_height_set)
request_natural_height = cached_size_request->natural_size;
else
request_natural_height = info->natural_height;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (min_height_p)
*min_height_p = request_min_height;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (natural_height_p)
*natural_height_p = request_natural_height;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
}
/**
* clutter_actor_get_allocation_box:
* @self: A #ClutterActor
* @box: (out): the function fills this in with the actor's allocation
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Gets the layout box an actor has been assigned. The allocation can
* only be assumed valid inside a paint() method; anywhere else, it
* may be out-of-date.
*
* An allocation does not incorporate the actor's scale or anchor point;
* those transformations do not affect layout, only rendering.
*
* <note>Do not call any of the clutter_actor_get_allocation_*() family
* of functions inside the implementation of the get_preferred_width()
* or get_preferred_height() virtual functions.</note>
*
* Since: 0.8
*/
void
clutter_actor_get_allocation_box (ClutterActor *self,
ClutterActorBox *box)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
/* XXX - if needs_allocation=TRUE, we can either 1) g_return_if_fail,
* which limits calling get_allocation to inside paint() basically; or
* we can 2) force a layout, which could be expensive if someone calls
* get_allocation somewhere silly; or we can 3) just return the latest
* value, allowing it to be out-of-date, and assume people know what
* they are doing.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* The least-surprises approach that keeps existing code working is
* likely to be 2). People can end up doing some inefficient things,
* though, and in general code that requires 2) is probably broken.
*/
/* this implements 2) */
if (G_UNLIKELY (self->priv->needs_allocation))
{
ClutterActor *stage = _clutter_actor_get_stage_internal (self);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* do not queue a relayout on an unparented actor */
if (stage)
_clutter_stage_maybe_relayout (stage);
}
/* commenting out the code above and just keeping this assigment
* implements 3)
*/
*box = self->priv->allocation;
}
/**
* clutter_actor_get_allocation_geometry:
* @self: A #ClutterActor
* @geom: (out): allocation geometry in pixels
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Gets the layout box an actor has been assigned. The allocation can
* only be assumed valid inside a paint() method; anywhere else, it
* may be out-of-date.
*
* An allocation does not incorporate the actor's scale or anchor point;
* those transformations do not affect layout, only rendering.
*
* The returned rectangle is in pixels.
*
* Since: 0.8
*/
void
clutter_actor_get_allocation_geometry (ClutterActor *self,
ClutterGeometry *geom)
{
ClutterActorBox box;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (geom != NULL);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_allocation_box (self, &box);
geom->x = CLUTTER_NEARBYINT (clutter_actor_box_get_x (&box));
geom->y = CLUTTER_NEARBYINT (clutter_actor_box_get_y (&box));
geom->width = CLUTTER_NEARBYINT (clutter_actor_box_get_width (&box));
geom->height = CLUTTER_NEARBYINT (clutter_actor_box_get_height (&box));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
}
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
static void
clutter_actor_update_constraints (ClutterActor *self,
ClutterActorBox *allocation)
{
ClutterActorPrivate *priv = self->priv;
const GList *constraints, *l;
if (priv->constraints == NULL)
return;
constraints = _clutter_meta_group_peek_metas (priv->constraints);
for (l = constraints; l != NULL; l = l->next)
{
ClutterConstraint *constraint = l->data;
ClutterActorMeta *meta = l->data;
if (clutter_actor_meta_get_enabled (meta))
{
_clutter_constraint_update_allocation (constraint,
self,
allocation);
}
}
}
/*< private >
* clutter_actor_adjust_allocation:
* @self: a #ClutterActor
* @allocation: (inout): the allocation to adjust
*
* Adjusts the passed allocation box taking into account the actor's
* layout information, like alignment, expansion, and margin.
*/
static void
clutter_actor_adjust_allocation (ClutterActor *self,
ClutterActorBox *allocation)
{
ClutterActorBox adj_allocation;
float alloc_width, alloc_height;
float min_width, min_height;
float nat_width, nat_height;
ClutterRequestMode req_mode;
adj_allocation = *allocation;
clutter_actor_box_get_size (allocation, &alloc_width, &alloc_height);
/* we want to hit the cache, so we use the public API */
req_mode = clutter_actor_get_request_mode (self);
if (req_mode == CLUTTER_REQUEST_HEIGHT_FOR_WIDTH)
{
clutter_actor_get_preferred_width (self, -1,
&min_width,
&nat_width);
clutter_actor_get_preferred_height (self, alloc_width,
&min_height,
&nat_height);
}
else if (req_mode == CLUTTER_REQUEST_WIDTH_FOR_HEIGHT)
{
clutter_actor_get_preferred_height (self, -1,
&min_height,
&nat_height);
clutter_actor_get_preferred_height (self, alloc_height,
&min_width,
&nat_width);
}
#ifdef CLUTTER_ENABLE_DEBUG
/* warn about underallocations */
if (_clutter_diagnostic_enabled () &&
(floorf (min_width - alloc_width) > 0 ||
floorf (min_height - alloc_height) > 0))
{
ClutterActor *parent = clutter_actor_get_parent (self);
/* the only actors that are allowed to be underallocated are the Stage,
* as it doesn't have an implicit size, and Actors that specifically
* told us that they want to opt-out from layout control mechanisms
* through the NO_LAYOUT escape hatch.
*/
if (parent != NULL &&
!(self->flags & CLUTTER_ACTOR_NO_LAYOUT) != 0)
{
g_warning (G_STRLOC ": The actor '%s' is getting an allocation "
"of %.2f x %.2f from its parent actor '%s', but its "
"requested minimum size is of %.2f x %.2f",
_clutter_actor_get_debug_name (self),
alloc_width, alloc_height,
_clutter_actor_get_debug_name (parent),
min_width, min_height);
}
}
#endif
clutter_actor_adjust_width (self,
&min_width,
&nat_width,
&adj_allocation.x1,
&adj_allocation.x2);
clutter_actor_adjust_height (self,
&min_height,
&nat_height,
&adj_allocation.y1,
&adj_allocation.y2);
/* we maintain the invariant that an allocation cannot be adjusted
* to be outside the parent-given box
*/
if (adj_allocation.x1 < allocation->x1 ||
adj_allocation.y1 < allocation->y1 ||
adj_allocation.x2 > allocation->x2 ||
adj_allocation.y2 > allocation->y2)
{
g_warning (G_STRLOC ": The actor '%s' tried to adjust its allocation "
"to { %.2f, %.2f, %.2f, %.2f }, which is outside of its "
"original allocation of { %.2f, %.2f, %.2f, %.2f }",
_clutter_actor_get_debug_name (self),
adj_allocation.x1, adj_allocation.y1,
adj_allocation.x2 - adj_allocation.x1,
adj_allocation.y2 - adj_allocation.y1,
allocation->x1, allocation->y1,
allocation->x2 - allocation->x1,
allocation->y2 - allocation->y1);
return;
}
*allocation = adj_allocation;
}
/**
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* clutter_actor_allocate:
* @self: A #ClutterActor
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* @box: new allocation of the actor, in parent-relative coordinates
* @flags: flags that control the allocation
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Called by the parent of an actor to assign the actor its size.
* Should never be called by applications (except when implementing
* a container or layout manager).
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Actors can know from their allocation box whether they have moved
* with respect to their parent actor. The @flags parameter describes
* additional information about the allocation, for instance whether
* the parent has moved with respect to the stage, for example because
* a grandparent's origin has moved.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.8
*/
void
clutter_actor_allocate (ClutterActor *self,
const ClutterActorBox *box,
ClutterAllocationFlags flags)
{
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
ClutterActorPrivate *priv;
ClutterActorClass *klass;
ClutterActorBox old_allocation, real_allocation;
gboolean origin_changed, child_moved, size_changed;
gboolean stage_allocation_changed;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (G_UNLIKELY (_clutter_actor_get_stage_internal (self) == NULL))
{
g_warning ("Spurious clutter_actor_allocate called for actor %p/%s "
"which isn't a descendent of the stage!\n",
self, _clutter_actor_get_debug_name (self));
return;
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv = self->priv;
old_allocation = priv->allocation;
real_allocation = *box;
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
/* constraints are allowed to modify the allocation only here; we do
* this prior to all the other checks so that we can bail out if the
* allocation did not change
*/
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
clutter_actor_update_constraints (self, &real_allocation);
/* adjust the allocation depending on the align/margin properties */
clutter_actor_adjust_allocation (self, &real_allocation);
if (real_allocation.x2 < real_allocation.x1 ||
real_allocation.y2 < real_allocation.y1)
{
g_warning (G_STRLOC ": Actor '%s' tried to allocate a size of %.2f x %.2f",
_clutter_actor_get_debug_name (self),
real_allocation.x2 - real_allocation.x1,
real_allocation.y2 - real_allocation.y1);
}
/* we allow 0-sized actors, but not negative-sized ones */
real_allocation.x2 = MAX (real_allocation.x2, real_allocation.x1);
real_allocation.y2 = MAX (real_allocation.y2, real_allocation.y1);
origin_changed = (flags & CLUTTER_ABSOLUTE_ORIGIN_CHANGED);
child_moved = (real_allocation.x1 != old_allocation.x1 ||
real_allocation.y1 != old_allocation.y1);
size_changed = (real_allocation.x2 != old_allocation.x2 ||
real_allocation.y2 != old_allocation.y2);
if (origin_changed || child_moved || size_changed)
stage_allocation_changed = TRUE;
else
stage_allocation_changed = FALSE;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* If we get an allocation "out of the blue"
* (we did not queue relayout), then we want to
* ignore it. But if we have needs_allocation set,
* we want to guarantee that allocate() virtual
* method is always called, i.e. that queue_relayout()
* always results in an allocate() invocation on
* an actor.
*
* The optimization here is to avoid re-allocating
* actors that did not queue relayout and were
* not moved.
*/
if (!priv->needs_allocation && !stage_allocation_changed)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
CLUTTER_NOTE (LAYOUT, "No allocation needed");
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
return;
}
/* When ABSOLUTE_ORIGIN_CHANGED is passed in to
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* clutter_actor_allocate(), it indicates whether the parent has its
* absolute origin moved; when passed in to ClutterActor::allocate()
* virtual method though, it indicates whether the child has its
* absolute origin moved. So we set it when child_moved is TRUE
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*/
if (child_moved)
flags |= CLUTTER_ABSOLUTE_ORIGIN_CHANGED;
CLUTTER_SET_PRIVATE_FLAGS (self, CLUTTER_IN_RELAYOUT);
klass = CLUTTER_ACTOR_GET_CLASS (self);
klass->allocate (self, &real_allocation, flags);
CLUTTER_UNSET_PRIVATE_FLAGS (self, CLUTTER_IN_RELAYOUT);
if (stage_allocation_changed)
clutter_actor_queue_redraw (self);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
}
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
/**
* clutter_actor_set_allocation:
* @self: a #ClutterActor
* @box: a #ClutterActorBox
* @flags: allocation flags
*
* Stores the allocation of @self as defined by @box.
*
* This function can only be called from within the implementation of
* the #ClutterActorClass.allocate() virtual function.
*
* The allocation should have been adjusted to take into account constraints,
* alignment, and margin properties. If you are implementing a #ClutterActor
* subclass that provides its own layout management policy for its children
* instead of using a #ClutterLayoutManager delegate, you should not call
* this function on the children of @self; instead, you should call
* clutter_actor_allocate(), which will adjust the allocation box for
* you.
*
* This function should only be used by subclasses of #ClutterActor
* that wish to store their allocation but cannot chain up to the
* parent's implementation; the default implementation of the
* #ClutterActorClass.allocate() virtual function will call this
* function.
*
* It is important to note that, while chaining up was the recommended
* behaviour for #ClutterActor subclasses prior to the introduction of
* this function, it is recommended to call clutter_actor_set_allocation()
* instead.
*
* If the #ClutterActor is using a #ClutterLayoutManager delegate object
* to handle the allocation of its children, this function will call
* the clutter_layout_manager_allocate() function only if the
* %CLUTTER_DELEGATE_LAYOUT flag is set on @flags, otherwise it is
* expected that the subclass will call clutter_layout_manager_allocate()
* by itself. For instance, the following code:
*
* |[
* static void
* my_actor_allocate (ClutterActor *actor,
* const ClutterActorBox *allocation,
* ClutterAllocationFlags flags)
* {
* ClutterActorBox new_alloc;
* ClutterAllocationFlags new_flags;
*
* adjust_allocation (allocation, &amp;new_alloc);
*
* new_flags = flags | CLUTTER_DELEGATE_LAYOUT;
*
* /&ast; this will use the layout manager set on the actor &ast;/
* clutter_actor_set_allocation (actor, &amp;new_alloc, new_flags);
* }
* ]|
*
* is equivalent to this:
*
* |[
* static void
* my_actor_allocate (ClutterActor *actor,
* const ClutterActorBox *allocation,
* ClutterAllocationFlags flags)
* {
* ClutterLayoutManager *layout;
* ClutterActorBox new_alloc;
*
* adjust_allocation (allocation, &amp;new_alloc);
*
* clutter_actor_set_allocation (actor, &amp;new_alloc, flags);
*
* layout = clutter_actor_get_layout_manager (actor);
* clutter_layout_manager_allocate (layout,
* CLUTTER_CONTAINER (actor),
* &amp;new_alloc,
* flags);
* }
* ]|
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
*
* Since: 1.10
*/
void
clutter_actor_set_allocation (ClutterActor *self,
const ClutterActorBox *box,
ClutterAllocationFlags flags)
{
ClutterActorPrivate *priv;
gboolean changed;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (box != NULL);
if (G_UNLIKELY (!CLUTTER_ACTOR_IN_RELAYOUT (self)))
{
g_critical (G_STRLOC ": The clutter_actor_set_allocation() function "
"can only be called from within the implementation of "
"the ClutterActor::allocate() virtual function.");
return;
}
priv = self->priv;
g_object_freeze_notify (G_OBJECT (self));
changed = clutter_actor_set_allocation_internal (self, box, flags);
/* we allocate our children before we notify changes in our geometry,
* so that people connecting to properties will be able to get valid
* data out of the sub-tree of the scene graph that has this actor at
* the root.
*/
clutter_actor_maybe_layout_children (self, box, flags);
actor: Maintain behaviour of old allocate() implementations The usual way to implement a container actor is to override the allocate() virtual function, chain up, and then allocate the actor's children. Clutter now has the ability to delegate layout management to ClutterLayoutManager directly; in the allocation, this is done by checking whether the actor has children, and then call clutter_layout_manager_allocate() from within the default implementation of the ClutterActor::allocate() vfunc. The same vfunc that everyone, has been chaining up to. Whoopsie. Well, we can check if there's a layout manager, and if it's NULL, we bail out. Except that there's a default layout manager, and it's the fixed layout manager, so that classes like Group and Stage work by default. Double whoopsie. The fix for this scenario is a bit nasty; we have to check if the actor class has overridden the allocate() vfunc or not, before actually looking at the layout manager. This means that classes that override the allocate() vfunc are expected to do everything that ClutterActor's default implementation does - which I think it's a fair requirement to have. For newly written code, though, it would probably be best if we just provided a function that does the right thing by default, and that you're supposed to be calling from within the allocate() vfunc implementation, if you ever chose to override it. This new function, clutter_actor_set_allocation(), should come with a warning the size of Texas, to avoid people thinking it's a way to override the whole "call allocate() on each child" mechanism. Plus, it should check if we're inside an allocation sequence, and bail out if not.
2012-01-19 12:40:32 +00:00
if (changed)
g_signal_emit (self, actor_signals[ALLOCATION_CHANGED], 0,
&priv->allocation,
priv->allocation_flags);
g_object_thaw_notify (G_OBJECT (self));
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* clutter_actor_set_geometry:
* @self: A #ClutterActor
* @geometry: A #ClutterGeometry
*
* Sets the actor's fixed position and forces its minimum and natural
* size, in pixels. This means the untransformed actor will have the
* given geometry. This is the same as calling clutter_actor_set_position()
* and clutter_actor_set_size().
*
* Deprecated: 1.10: Use clutter_actor_set_position() and
* clutter_actor_set_size() instead.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*/
void
clutter_actor_set_geometry (ClutterActor *self,
const ClutterGeometry *geometry)
{
g_object_freeze_notify (G_OBJECT (self));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_set_position (self, geometry->x, geometry->y);
clutter_actor_set_size (self, geometry->width, geometry->height);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_thaw_notify (G_OBJECT (self));
}
/**
* clutter_actor_get_geometry:
* @self: A #ClutterActor
* @geometry: (out caller-allocates): A location to store actors #ClutterGeometry
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Gets the size and position of an actor relative to its parent
* actor. This is the same as calling clutter_actor_get_position() and
* clutter_actor_get_size(). It tries to "do what you mean" and get the
* requested size and position if the actor's allocation is invalid.
*
* Deprecated: 1.10: Use clutter_actor_get_position() and
* clutter_actor_get_size(), or clutter_actor_get_allocation_geometry()
* instead.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*/
void
clutter_actor_get_geometry (ClutterActor *self,
ClutterGeometry *geometry)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat x, y, width, height;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_return_if_fail (geometry != NULL);
clutter_actor_get_position (self, &x, &y);
clutter_actor_get_size (self, &width, &height);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
geometry->x = (int) x;
geometry->y = (int) y;
geometry->width = (int) width;
geometry->height = (int) height;
}
/**
* clutter_actor_set_position:
* @self: A #ClutterActor
* @x: New left position of actor in pixels.
* @y: New top position of actor in pixels.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Sets the actor's fixed position in pixels relative to any parent
* actor.
*
* If a layout manager is in use, this position will override the
* layout manager and force a fixed position.
*/
void
clutter_actor_set_position (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat x,
gfloat y)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_freeze_notify (G_OBJECT (self));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_set_x (self, x);
clutter_actor_set_y (self, y);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_thaw_notify (G_OBJECT (self));
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
* clutter_actor_get_fixed_position_set:
* @self: A #ClutterActor
*
* Checks whether an actor has a fixed position set (and will thus be
* unaffected by any layout manager).
*
* Return value: %TRUE if the fixed position is set on the actor
*
* Since: 0.8
*/
gboolean
clutter_actor_get_fixed_position_set (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
return self->priv->position_set;
}
/**
* clutter_actor_set_fixed_position_set:
* @self: A #ClutterActor
* @is_set: whether to use fixed position
*
* Sets whether an actor has a fixed position set (and will thus be
* unaffected by any layout manager).
*
* Since: 0.8
*/
void
clutter_actor_set_fixed_position_set (ClutterActor *self,
gboolean is_set)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (self->priv->position_set == (is_set != FALSE))
return;
self->priv->position_set = is_set != FALSE;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_FIXED_POSITION_SET]);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_move_by:
* @self: A #ClutterActor
* @dx: Distance to move Actor on X axis.
* @dy: Distance to move Actor on Y axis.
*
* Moves an actor by the specified distance relative to its current
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* position in pixels.
*
* This function modifies the fixed position of an actor and thus removes
* it from any layout management. Another way to move an actor is with an
* anchor point, see clutter_actor_set_anchor_point().
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Since: 0.2
*/
void
clutter_actor_move_by (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat dx,
gfloat dy)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
const ClutterLayoutInfo *info;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat x, y;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
info = _clutter_actor_get_layout_info_or_defaults (self);
x = info->fixed_x;
y = info->fixed_y;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_set_position (self, x + dx, y + dy);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
}
static void
clutter_actor_set_min_width (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat min_width)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
ClutterActorPrivate *priv = self->priv;
ClutterActorBox old = { 0, };
ClutterLayoutInfo *info;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* if we are setting the size on a top-level actor and the
* backend only supports static top-levels (e.g. framebuffers)
* then we ignore the passed value and we override it with
* the stage implementation's preferred size.
*/
if (CLUTTER_ACTOR_IS_TOPLEVEL (self) &&
clutter_feature_available (CLUTTER_FEATURE_STAGE_STATIC))
return;
info = _clutter_actor_get_layout_info (self);
if (priv->min_width_set && min_width == info->min_width)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
return;
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_store_old_geometry (self, &old);
info->min_width = min_width;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MIN_WIDTH]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_set_min_width_set (self, TRUE);
clutter_actor_notify_if_geometry_changed (self, &old);
g_object_thaw_notify (G_OBJECT (self));
clutter_actor_queue_relayout (self);
}
static void
clutter_actor_set_min_height (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat min_height)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
ClutterActorPrivate *priv = self->priv;
ClutterActorBox old = { 0, };
ClutterLayoutInfo *info;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* if we are setting the size on a top-level actor and the
* backend only supports static top-levels (e.g. framebuffers)
* then we ignore the passed value and we override it with
* the stage implementation's preferred size.
*/
if (CLUTTER_ACTOR_IS_TOPLEVEL (self) &&
clutter_feature_available (CLUTTER_FEATURE_STAGE_STATIC))
return;
info = _clutter_actor_get_layout_info (self);
if (priv->min_height_set && min_height == info->min_height)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
return;
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_store_old_geometry (self, &old);
info->min_height = min_height;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MIN_HEIGHT]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_set_min_height_set (self, TRUE);
clutter_actor_notify_if_geometry_changed (self, &old);
g_object_thaw_notify (G_OBJECT (self));
clutter_actor_queue_relayout (self);
}
static void
clutter_actor_set_natural_width (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat natural_width)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
ClutterActorPrivate *priv = self->priv;
ClutterActorBox old = { 0, };
ClutterLayoutInfo *info;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* if we are setting the size on a top-level actor and the
* backend only supports static top-levels (e.g. framebuffers)
* then we ignore the passed value and we override it with
* the stage implementation's preferred size.
*/
if (CLUTTER_ACTOR_IS_TOPLEVEL (self) &&
clutter_feature_available (CLUTTER_FEATURE_STAGE_STATIC))
return;
info = _clutter_actor_get_layout_info (self);
if (priv->natural_width_set && natural_width == info->natural_width)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
return;
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_store_old_geometry (self, &old);
info->natural_width = natural_width;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_NATURAL_WIDTH]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_set_natural_width_set (self, TRUE);
clutter_actor_notify_if_geometry_changed (self, &old);
g_object_thaw_notify (G_OBJECT (self));
clutter_actor_queue_relayout (self);
}
static void
clutter_actor_set_natural_height (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat natural_height)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
ClutterActorPrivate *priv = self->priv;
ClutterActorBox old = { 0, };
ClutterLayoutInfo *info;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* if we are setting the size on a top-level actor and the
* backend only supports static top-levels (e.g. framebuffers)
* then we ignore the passed value and we override it with
* the stage implementation's preferred size.
*/
if (CLUTTER_ACTOR_IS_TOPLEVEL (self) &&
clutter_feature_available (CLUTTER_FEATURE_STAGE_STATIC))
return;
info = _clutter_actor_get_layout_info (self);
if (priv->natural_height_set && natural_height == info->natural_height)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
return;
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_store_old_geometry (self, &old);
info->natural_height = natural_height;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_NATURAL_HEIGHT]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_set_natural_height_set (self, TRUE);
clutter_actor_notify_if_geometry_changed (self, &old);
g_object_thaw_notify (G_OBJECT (self));
clutter_actor_queue_relayout (self);
}
static void
clutter_actor_set_min_width_set (ClutterActor *self,
gboolean use_min_width)
{
ClutterActorPrivate *priv = self->priv;
ClutterActorBox old = { 0, };
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (priv->min_width_set == (use_min_width != FALSE))
return;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_store_old_geometry (self, &old);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->min_width_set = use_min_width != FALSE;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MIN_WIDTH_SET]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_notify_if_geometry_changed (self, &old);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_queue_relayout (self);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
static void
clutter_actor_set_min_height_set (ClutterActor *self,
gboolean use_min_height)
{
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
ClutterActorPrivate *priv = self->priv;
ClutterActorBox old = { 0, };
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (priv->min_height_set == (use_min_height != FALSE))
return;
clutter_actor_store_old_geometry (self, &old);
priv->min_height_set = use_min_height != FALSE;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MIN_HEIGHT_SET]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_notify_if_geometry_changed (self, &old);
clutter_actor_queue_relayout (self);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
static void
clutter_actor_set_natural_width_set (ClutterActor *self,
gboolean use_natural_width)
{
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
ClutterActorPrivate *priv = self->priv;
ClutterActorBox old = { 0, };
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (priv->natural_width_set == (use_natural_width != FALSE))
return;
clutter_actor_store_old_geometry (self, &old);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->natural_width_set = use_natural_width != FALSE;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_NATURAL_WIDTH_SET]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_notify_if_geometry_changed (self, &old);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_queue_relayout (self);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
static void
clutter_actor_set_natural_height_set (ClutterActor *self,
gboolean use_natural_height)
{
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
ClutterActorPrivate *priv = self->priv;
ClutterActorBox old = { 0, };
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (priv->natural_height_set == (use_natural_height != FALSE))
return;
clutter_actor_store_old_geometry (self, &old);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->natural_height_set = use_natural_height != FALSE;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_NATURAL_HEIGHT_SET]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_notify_if_geometry_changed (self, &old);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_set_request_mode:
* @self: a #ClutterActor
* @mode: the request mode
*
* Sets the geometry request mode of @self.
*
* The @mode determines the order for invoking
* clutter_actor_get_preferred_width() and
* clutter_actor_get_preferred_height()
*
* Since: 1.2
*/
void
clutter_actor_set_request_mode (ClutterActor *self,
ClutterRequestMode mode)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (priv->request_mode == mode)
return;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->request_mode = mode;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
priv->needs_width_request = TRUE;
priv->needs_height_request = TRUE;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_REQUEST_MODE]);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_get_request_mode:
* @self: a #ClutterActor
*
* Retrieves the geometry request mode of @self
*
* Return value: the request mode for the actor
*
* Since: 1.2
*/
ClutterRequestMode
clutter_actor_get_request_mode (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self),
CLUTTER_REQUEST_HEIGHT_FOR_WIDTH);
return self->priv->request_mode;
}
/* variant of set_width() without checks and without notification
* freeze+thaw, for internal usage only
*/
static inline void
clutter_actor_set_width_internal (ClutterActor *self,
gfloat width)
{
if (width >= 0)
{
/* the Stage will use the :min-width to control the minimum
* width to be resized to, so we should not be setting it
* along with the :natural-width
*/
if (!CLUTTER_ACTOR_IS_TOPLEVEL (self))
clutter_actor_set_min_width (self, width);
clutter_actor_set_natural_width (self, width);
}
else
{
/* we only unset the :natural-width for the Stage */
if (!CLUTTER_ACTOR_IS_TOPLEVEL (self))
clutter_actor_set_min_width_set (self, FALSE);
clutter_actor_set_natural_width_set (self, FALSE);
}
}
/* variant of set_height() without checks and without notification
* freeze+thaw, for internal usage only
*/
static inline void
clutter_actor_set_height_internal (ClutterActor *self,
gfloat height)
{
if (height >= 0)
{
/* see the comment above in set_width_internal() */
if (!CLUTTER_ACTOR_IS_TOPLEVEL (self))
clutter_actor_set_min_height (self, height);
clutter_actor_set_natural_height (self, height);
}
else
{
/* see the comment above in set_width_internal() */
if (!CLUTTER_ACTOR_IS_TOPLEVEL (self))
clutter_actor_set_min_height_set (self, FALSE);
clutter_actor_set_natural_height_set (self, FALSE);
}
}
/**
* clutter_actor_set_size:
* @self: A #ClutterActor
* @width: New width of actor in pixels, or -1
* @height: New height of actor in pixels, or -1
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Sets the actor's size request in pixels. This overrides any
* "normal" size request the actor would have. For example
* a text actor might normally request the size of the text;
* this function would force a specific size instead.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* If @width and/or @height are -1 the actor will use its
* "normal" size request instead of overriding it, i.e.
* you can "unset" the size with -1.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* This function sets or unsets both the minimum and natural size.
*/
void
clutter_actor_set_size (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat width,
gfloat height)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_set_width_internal (self, width);
clutter_actor_set_height_internal (self, height);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_thaw_notify (G_OBJECT (self));
}
/**
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
* clutter_actor_get_size:
* @self: A #ClutterActor
2009-07-02 13:21:16 +00:00
* @width: (out) (allow-none): return location for the width, or %NULL.
* @height: (out) (allow-none): return location for the height, or %NULL.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* This function tries to "do what you mean" and return
* the size an actor will have. If the actor has a valid
* allocation, the allocation will be returned; otherwise,
* the actors natural size request will be returned.
*
* If you care whether you get the request vs. the allocation, you
* should probably call a different function like
* clutter_actor_get_allocation_box() or
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* clutter_actor_get_preferred_width().
*
* Since: 0.2
*/
void
clutter_actor_get_size (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat *width,
gfloat *height)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (width)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*width = clutter_actor_get_width (self);
if (height)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*height = clutter_actor_get_height (self);
}
/**
* clutter_actor_get_position:
* @self: a #ClutterActor
2009-07-02 13:21:16 +00:00
* @x: (out) (allow-none): return location for the X coordinate, or %NULL
* @y: (out) (allow-none): return location for the Y coordinate, or %NULL
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* This function tries to "do what you mean" and tell you where the
* actor is, prior to any transformations. Retrieves the fixed
* position of an actor in pixels, if one has been set; otherwise, if
* the allocation is valid, returns the actor's allocated position;
* otherwise, returns 0,0.
*
* The returned position is in pixels.
*
* Since: 0.6
*/
void
clutter_actor_get_position (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat *x,
gfloat *y)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (x)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*x = clutter_actor_get_x (self);
if (y)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*y = clutter_actor_get_y (self);
}
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_get_transformed_position:
* @self: A #ClutterActor
2009-07-02 13:21:16 +00:00
* @x: (out) (allow-none): return location for the X coordinate, or %NULL
* @y: (out) (allow-none): return location for the Y coordinate, or %NULL
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Gets the absolute position of an actor, in pixels relative to the stage.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Since: 0.8
*/
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_get_transformed_position (ClutterActor *self,
gfloat *x,
gfloat *y)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
ClutterVertex v1;
ClutterVertex v2;
v1.x = v1.y = v1.z = 0;
clutter_actor_apply_transform_to_point (self, &v1, &v2);
if (x)
*x = v2.x;
if (y)
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
*y = v2.y;
2007-05-18 07:30:06 +00:00
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_get_transformed_size:
* @self: A #ClutterActor
2009-07-02 13:21:16 +00:00
* @width: (out) (allow-none): return location for the width, or %NULL
* @height: (out) (allow-none): return location for the height, or %NULL
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Gets the absolute size of an actor in pixels, taking into account the
* scaling factors.
*
* If the actor has a valid allocation, the allocated size will be used.
* If the actor has not a valid allocation then the preferred size will
* be transformed and returned.
*
* If you want the transformed allocation, see
* clutter_actor_get_abs_allocation_vertices() instead.
*
* <note>When the actor (or one of its ancestors) is rotated around the
* X or Y axis, it no longer appears as on the stage as a rectangle, but
* as a generic quadrangle; in that case this function returns the size
* of the smallest rectangle that encapsulates the entire quad. Please
* note that in this case no assumptions can be made about the relative
* position of this envelope to the absolute position of the actor, as
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* returned by clutter_actor_get_transformed_position(); if you need this
* information, you need to use clutter_actor_get_abs_allocation_vertices()
* to get the coords of the actual quadrangle.</note>
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Since: 0.8
*/
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_get_transformed_size (ClutterActor *self,
gfloat *width,
gfloat *height)
{
ClutterActorPrivate *priv;
ClutterVertex v[4];
gfloat x_min, x_max, y_min, y_max;
gint i;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
/* if the actor hasn't been allocated yet, get the preferred
* size and transform that
*/
if (priv->needs_allocation)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat natural_width, natural_height;
ClutterActorBox box;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
/* Make a fake allocation to transform.
*
* NB: _clutter_actor_transform_and_project_box expects a box in
* the actor's coordinate space... */
box.x1 = 0;
box.y1 = 0;
natural_width = natural_height = 0;
clutter_actor_get_preferred_size (self, NULL, NULL,
&natural_width,
&natural_height);
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
box.x2 = natural_width;
box.y2 = natural_height;
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
_clutter_actor_transform_and_project_box (self, &box, v);
}
else
clutter_actor_get_abs_allocation_vertices (self, v);
x_min = x_max = v[0].x;
y_min = y_max = v[0].y;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
for (i = 1; i < G_N_ELEMENTS (v); ++i)
{
if (v[i].x < x_min)
x_min = v[i].x;
if (v[i].x > x_max)
x_max = v[i].x;
if (v[i].y < y_min)
y_min = v[i].y;
if (v[i].y > y_max)
y_max = v[i].y;
2007-05-18 07:30:06 +00:00
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (width)
*width = x_max - x_min;
if (height)
*height = y_max - y_min;
2007-05-18 07:30:06 +00:00
}
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_get_width:
* @self: A #ClutterActor
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Retrieves the width of a #ClutterActor.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* If the actor has a valid allocation, this function will return the
* width of the allocated area given to the actor.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* If the actor does not have a valid allocation, this function will
* return the actor's natural width, that is the preferred width of
* the actor.
*
* If you care whether you get the preferred width or the width that
* has been assigned to the actor, you should probably call a different
* function like clutter_actor_get_allocation_box() to retrieve the
* allocated size or clutter_actor_get_preferred_width() to retrieve the
* preferred width.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* If an actor has a fixed width, for instance a width that has been
* assigned using clutter_actor_set_width(), the width returned will
* be the same value.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Return value: the width of the actor, in pixels
*/
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat
clutter_actor_get_width (ClutterActor *self)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
ClutterActorPrivate *priv;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
priv = self->priv;
if (priv->needs_allocation)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat natural_width = 0;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (self->priv->request_mode == CLUTTER_REQUEST_HEIGHT_FOR_WIDTH)
clutter_actor_get_preferred_width (self, -1, NULL, &natural_width);
else
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat natural_height = 0;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_preferred_height (self, -1, NULL, &natural_height);
clutter_actor_get_preferred_width (self, natural_height,
NULL,
&natural_width);
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
return natural_width;
}
else
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
return priv->allocation.x2 - priv->allocation.x1;
}
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_get_height:
* @self: A #ClutterActor
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Retrieves the height of a #ClutterActor.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* If the actor has a valid allocation, this function will return the
* height of the allocated area given to the actor.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* If the actor does not have a valid allocation, this function will
* return the actor's natural height, that is the preferred height of
* the actor.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* If you care whether you get the preferred height or the height that
* has been assigned to the actor, you should probably call a different
* function like clutter_actor_get_allocation_box() to retrieve the
* allocated size or clutter_actor_get_preferred_height() to retrieve the
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* preferred height.
*
* If an actor has a fixed height, for instance a height that has been
* assigned using clutter_actor_set_height(), the height returned will
* be the same value.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Return value: the height of the actor, in pixels
*/
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat
clutter_actor_get_height (ClutterActor *self)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
ClutterActorPrivate *priv;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
priv = self->priv;
if (priv->needs_allocation)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat natural_height = 0;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (priv->request_mode == CLUTTER_REQUEST_HEIGHT_FOR_WIDTH)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat natural_width = 0;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_preferred_width (self, -1, NULL, &natural_width);
clutter_actor_get_preferred_height (self, natural_width,
NULL, &natural_height);
}
else
clutter_actor_get_preferred_height (self, -1, NULL, &natural_height);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
return natural_height;
}
else
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
return priv->allocation.y2 - priv->allocation.y1;
}
/**
* clutter_actor_set_width:
* @self: A #ClutterActor
* @width: Requested new width for the actor, in pixels, or -1
*
* Forces a width on an actor, causing the actor's preferred width
* and height (if any) to be ignored.
*
* If @width is -1 the actor will use its preferred width request
* instead of overriding it, i.e. you can "unset" the width with -1.
*
* This function sets both the minimum and natural size of the actor.
*
* since: 0.2
*/
void
clutter_actor_set_width (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat width)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_set_width_internal (self, width);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_thaw_notify (G_OBJECT (self));
}
/**
* clutter_actor_set_height:
* @self: A #ClutterActor
* @height: Requested new height for the actor, in pixels, or -1
*
* Forces a height on an actor, causing the actor's preferred width
* and height (if any) to be ignored.
*
* If @height is -1 the actor will use its preferred height instead of
* overriding it, i.e. you can "unset" the height with -1.
*
* This function sets both the minimum and natural size of the actor.
*
* since: 0.2
*/
void
clutter_actor_set_height (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat height)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_set_height_internal (self, height);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_object_thaw_notify (G_OBJECT (self));
}
/**
* clutter_actor_set_x:
* @self: a #ClutterActor
* @x: the actor's position on the X axis
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Sets the actor's X coordinate, relative to its parent, in pixels.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Overrides any layout manager and forces a fixed position for
* the actor.
*
* Since: 0.6
*/
void
clutter_actor_set_x (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat x)
{
ClutterActorBox old = { 0, };
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
ClutterActorPrivate *priv;
ClutterLayoutInfo *info;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
priv = self->priv;
info = _clutter_actor_get_layout_info (self);
if (priv->position_set && info->fixed_x == x)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
return;
clutter_actor_store_old_geometry (self, &old);
info->fixed_x = x;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_set_fixed_position_set (self, TRUE);
clutter_actor_notify_if_geometry_changed (self, &old);
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_set_y:
* @self: a #ClutterActor
* @y: the actor's position on the Y axis
*
* Sets the actor's Y coordinate, relative to its parent, in pixels.#
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Overrides any layout manager and forces a fixed position for
* the actor.
*
* Since: 0.6
*/
void
clutter_actor_set_y (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat y)
{
ClutterActorBox old = { 0, };
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
ClutterActorPrivate *priv;
ClutterLayoutInfo *info;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
priv = self->priv;
info = _clutter_actor_get_layout_info (self);
if (priv->position_set && info->fixed_y == y)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
return;
clutter_actor_store_old_geometry (self, &old);
info->fixed_y = y;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_set_fixed_position_set (self, TRUE);
clutter_actor_notify_if_geometry_changed (self, &old);
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_get_x:
* @self: A #ClutterActor
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Retrieves the X coordinate of a #ClutterActor.
*
* This function tries to "do what you mean", by returning the
* correct value depending on the actor's state.
*
* If the actor has a valid allocation, this function will return
* the X coordinate of the origin of the allocation box.
*
* If the actor has any fixed coordinate set using clutter_actor_set_x(),
* clutter_actor_set_position() or clutter_actor_set_geometry(), this
* function will return that coordinate.
*
* If both the allocation and a fixed position are missing, this function
* will return 0.
*
* Return value: the X coordinate, in pixels, ignoring any
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* transformation (i.e. scaling, rotation)
*/
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat
clutter_actor_get_x (ClutterActor *self)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
ClutterActorPrivate *priv;
2007-05-18 07:30:06 +00:00
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
priv = self->priv;
if (priv->needs_allocation)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (priv->position_set)
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (self);
return info->fixed_x;
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
else
return 0;
}
else
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
return priv->allocation.x1;
}
/**
* clutter_actor_get_y:
* @self: A #ClutterActor
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Retrieves the Y coordinate of a #ClutterActor.
*
* This function tries to "do what you mean", by returning the
* correct value depending on the actor's state.
*
* If the actor has a valid allocation, this function will return
* the Y coordinate of the origin of the allocation box.
*
* If the actor has any fixed coordinate set using clutter_actor_set_y(),
* clutter_actor_set_position() or clutter_actor_set_geometry(), this
* function will return that coordinate.
*
* If both the allocation and a fixed position are missing, this function
* will return 0.
*
* Return value: the Y coordinate, in pixels, ignoring any
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* transformation (i.e. scaling, rotation)
*/
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat
clutter_actor_get_y (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
priv = self->priv;
if (priv->needs_allocation)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
if (priv->position_set)
{
const ClutterLayoutInfo *info;
info = _clutter_actor_get_layout_info_or_defaults (self);
return info->fixed_y;
}
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
else
return 0;
}
else
return priv->allocation.y1;
}
/**
* clutter_actor_set_scale:
* @self: A #ClutterActor
* @scale_x: double factor to scale actor by horizontally.
* @scale_y: double factor to scale actor by vertically.
*
* Scales an actor with the given factors. The scaling is relative to
* the scale center and the anchor point. The scale center is
* unchanged by this function and defaults to 0,0.
*
* Since: 0.2
*/
void
clutter_actor_set_scale (ClutterActor *self,
gdouble scale_x,
gdouble scale_y)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_set_scale_factor (self, CLUTTER_X_AXIS, scale_x);
clutter_actor_set_scale_factor (self, CLUTTER_Y_AXIS, scale_y);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_object_thaw_notify (G_OBJECT (self));
}
/**
* clutter_actor_set_scale_full:
* @self: A #ClutterActor
* @scale_x: double factor to scale actor by horizontally.
* @scale_y: double factor to scale actor by vertically.
* @center_x: X coordinate of the center of the scale.
* @center_y: Y coordinate of the center of the scale
*
* Scales an actor with the given factors around the given center
* point. The center point is specified in pixels relative to the
* anchor point (usually the top left corner of the actor).
*
* Since: 1.0
*/
void
clutter_actor_set_scale_full (ClutterActor *self,
gdouble scale_x,
gdouble scale_y,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat center_x,
gfloat center_y)
{
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_set_scale_factor (self, CLUTTER_X_AXIS, scale_x);
clutter_actor_set_scale_factor (self, CLUTTER_Y_AXIS, scale_y);
clutter_actor_set_scale_center (self, CLUTTER_X_AXIS, center_x);
clutter_actor_set_scale_center (self, CLUTTER_Y_AXIS, center_y);
g_object_thaw_notify (G_OBJECT (self));
}
/**
* clutter_actor_set_scale_with_gravity:
* @self: A #ClutterActor
* @scale_x: double factor to scale actor by horizontally.
* @scale_y: double factor to scale actor by vertically.
* @gravity: the location of the scale center expressed as a compass
* direction.
*
* Scales an actor with the given factors around the given
* center point. The center point is specified as one of the compass
* directions in #ClutterGravity. For example, setting it to north
* will cause the top of the actor to remain unchanged and the rest of
* the actor to expand left, right and downwards.
*
* Since: 1.0
*/
void
clutter_actor_set_scale_with_gravity (ClutterActor *self,
gdouble scale_x,
gdouble scale_y,
ClutterGravity gravity)
{
ClutterTransformInfo *info;
GObject *obj;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
obj = G_OBJECT (self);
g_object_freeze_notify (obj);
info = _clutter_actor_get_transform_info (self);
info->scale_x = scale_x;
info->scale_y = scale_y;
if (gravity == CLUTTER_GRAVITY_NONE)
clutter_anchor_coord_set_units (&info->scale_center, 0, 0, 0);
else
clutter_anchor_coord_set_gravity (&info->scale_center, gravity);
self->priv->transform_valid = FALSE;
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_X]);
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_Y]);
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_CENTER_X]);
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_CENTER_Y]);
g_object_notify_by_pspec (obj, obj_props[PROP_SCALE_GRAVITY]);
clutter_actor_queue_redraw (self);
g_object_thaw_notify (obj);
}
/**
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
* clutter_actor_get_scale:
* @self: A #ClutterActor
2009-07-02 13:21:16 +00:00
* @scale_x: (out) (allow-none): Location to store horizonal
* scale factor, or %NULL.
* @scale_y: (out) (allow-none): Location to store vertical
* scale factor, or %NULL.
*
2009-07-02 13:21:16 +00:00
* Retrieves an actors scale factors.
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
*
* Since: 0.2
*/
void
clutter_actor_get_scale (ClutterActor *self,
gdouble *scale_x,
gdouble *scale_y)
{
const ClutterTransformInfo *info;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
g_return_if_fail (CLUTTER_IS_ACTOR (self));
info = _clutter_actor_get_transform_info_or_defaults (self);
if (scale_x)
*scale_x = info->scale_x;
if (scale_y)
*scale_y = info->scale_y;
}
/**
* clutter_actor_get_scale_center:
* @self: A #ClutterActor
2009-07-02 13:21:16 +00:00
* @center_x: (out) (allow-none): Location to store the X position
* of the scale center, or %NULL.
* @center_y: (out) (allow-none): Location to store the Y position
* of the scale center, or %NULL.
*
* Retrieves the scale center coordinate in pixels relative to the top
* left corner of the actor. If the scale center was specified using a
* #ClutterGravity this will calculate the pixel offset using the
* current size of the actor.
*
* Since: 1.0
*/
void
clutter_actor_get_scale_center (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat *center_x,
gfloat *center_y)
{
const ClutterTransformInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
info = _clutter_actor_get_transform_info_or_defaults (self);
clutter_anchor_coord_get_units (self, &info->scale_center,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
center_x,
center_y,
NULL);
}
/**
* clutter_actor_get_scale_gravity:
* @self: A #ClutterActor
*
* Retrieves the scale center as a compass direction. If the scale
* center was specified in pixels or units this will return
* %CLUTTER_GRAVITY_NONE.
*
* Return value: the scale gravity
*
* Since: 1.0
*/
ClutterGravity
clutter_actor_get_scale_gravity (ClutterActor *self)
{
const ClutterTransformInfo *info;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), CLUTTER_GRAVITY_NONE);
info = _clutter_actor_get_transform_info_or_defaults (self);
return clutter_anchor_coord_get_gravity (&info->scale_center);
}
/**
* clutter_actor_set_opacity:
* @self: A #ClutterActor
* @opacity: New opacity value for the actor.
*
* Sets the actor's opacity, with zero being completely transparent and
* 255 (0xff) being fully opaque.
*/
void
clutter_actor_set_opacity (ClutterActor *self,
guint8 opacity)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
priv = self->priv;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (priv->opacity != opacity)
{
priv->opacity = opacity;
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
/* Queue a redraw from the flatten effect so that it can use
its cached image if available instead of having to redraw the
actual actor. If it doesn't end up using the FBO then the
effect is still able to continue the paint anyway. If there
is no flatten effect yet then this is equivalent to queueing
a full redraw */
_clutter_actor_queue_redraw_full (self,
0, /* flags */
NULL, /* clip */
priv->flatten_effect);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_OPACITY]);
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
}
}
/*
* clutter_actor_get_paint_opacity_internal:
* @self: a #ClutterActor
*
* Retrieves the absolute opacity of the actor, as it appears on the stage
*
* This function does not do type checks
*
* Return value: the absolute opacity of the actor
*/
static guint8
clutter_actor_get_paint_opacity_internal (ClutterActor *self)
{
ClutterActorPrivate *priv = self->priv;
ClutterActor *parent;
/* override the top-level opacity to always be 255; even in
* case of ClutterStage:use-alpha being TRUE we want the rest
* of the scene to be painted
*/
if (CLUTTER_ACTOR_IS_TOPLEVEL (self))
return 255;
if (priv->opacity_override >= 0)
return priv->opacity_override;
parent = priv->parent;
/* Factor in the actual actors opacity with parents */
if (parent != NULL)
{
guint8 opacity = clutter_actor_get_paint_opacity_internal (parent);
if (opacity != 0xff)
return (opacity * priv->opacity) / 0xff;
}
return priv->opacity;
}
/**
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* clutter_actor_get_paint_opacity:
* @self: A #ClutterActor
*
* Retrieves the absolute opacity of the actor, as it appears on the stage.
*
* This function traverses the hierarchy chain and composites the opacity of
* the actor with that of its parents.
*
* This function is intended for subclasses to use in the paint virtual
* function, to paint themselves with the correct opacity.
*
* Return value: The actor opacity value.
*
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* Since: 0.8
*/
guint8
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_paint_opacity (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
return clutter_actor_get_paint_opacity_internal (self);
}
/**
* clutter_actor_get_opacity:
* @self: a #ClutterActor
*
* Retrieves the opacity value of an actor, as set by
* clutter_actor_set_opacity().
*
* For retrieving the absolute opacity of the actor inside a paint
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
* virtual function, see clutter_actor_get_paint_opacity().
*
* Return value: the opacity of the actor
*/
guint8
clutter_actor_get_opacity (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
return self->priv->opacity;
}
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
/**
* clutter_actor_set_offscreen_redirect:
* @self: A #ClutterActor
* @redirect: New offscreen redirect flags for the actor.
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
*
* Defines the circumstances where the actor should be redirected into
* an offscreen image. The offscreen image is used to flatten the
* actor into a single image while painting for two main reasons.
* Firstly, when the actor is painted a second time without any of its
* contents changing it can simply repaint the cached image without
* descending further down the actor hierarchy. Secondly, it will make
* the opacity look correct even if there are overlapping primitives
* in the actor.
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
*
* Caching the actor could in some cases be a performance win and in
* some cases be a performance lose so it is important to determine
* which value is right for an actor before modifying this value. For
* example, there is never any reason to flatten an actor that is just
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
* a single texture (such as a #ClutterTexture) because it is
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
* effectively already cached in an image so the offscreen would be
* redundant. Also if the actor contains primitives that are far apart
* with a large transparent area in the middle (such as a large
* CluterGroup with a small actor in the top left and a small actor in
* the bottom right) then the cached image will contain the entire
* image of the large area and the paint will waste time blending all
* of the transparent pixels in the middle.
*
* The default method of implementing opacity on a container simply
* forwards on the opacity to all of the children. If the children are
* overlapping then it will appear as if they are two separate glassy
* objects and there will be a break in the color where they
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
* overlap. By redirecting to an offscreen buffer it will be as if the
* two opaque objects are combined into one and then made transparent
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
* which is usually what is expected.
*
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
* The image below demonstrates the difference between redirecting and
* not. The image shows two Clutter groups, each containing a red and
* a green rectangle which overlap. The opacity on the group is set to
* 128 (which is 50%). When the offscreen redirect is not used, the
* red rectangle can be seen through the blue rectangle as if the two
* rectangles were separately transparent. When the redirect is used
* the group as a whole is transparent instead so the red rectangle is
* not visible where they overlap.
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
*
* <figure id="offscreen-redirect">
* <title>Sample of using an offscreen redirect for transparency</title>
* <graphic fileref="offscreen-redirect.png" format="PNG"/>
* </figure>
*
* The default value for this property is 0, so we effectively will
* never redirect an actor offscreen by default. This means that there
* are times that transparent actors may look glassy as described
* above. The reason this is the default is because there is a
* performance trade off between quality and performance here. In many
* cases the default form of glassy opacity looks good enough, but if
* it's not you will need to set the
* %CLUTTER_OFFSCREEN_REDIRECT_AUTOMATIC_FOR_OPACITY flag to enable
* redirection for opacity.
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
*
* Custom actors that don't contain any overlapping primitives are
* recommended to override the has_overlaps() virtual to return %FALSE
* for maximum efficiency.
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
*
* Since: 1.8
*/
void
clutter_actor_set_offscreen_redirect (ClutterActor *self,
ClutterOffscreenRedirect redirect)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
if (priv->offscreen_redirect != redirect)
{
priv->offscreen_redirect = redirect;
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
/* Queue a redraw from the effect so that it can use its cached
image if available instead of having to redraw the actual
actor. If it doesn't end up using the FBO then the effect is
still able to continue the paint anyway. If there is no
effect then this is equivalent to queuing a full redraw */
_clutter_actor_queue_redraw_full (self,
0, /* flags */
NULL, /* clip */
priv->flatten_effect);
clutter-actor: Add an 'offscreen-redirect' property This adds a property which can be used to redirect the actor through an FBO before painting so that it becomes flattened in an image. The image can be used as a cache to avoid having to repaint the actor if something unrelated in the scene changes. It can also be used to implement correct opacity even if the actor has overlapping primitives. The property is an enum that takes three values: CLUTTER_OFFSCREEN_REDIRECT_NEVER: The default behaviour which is to never flatten the actor. CLUTTER_OFFSCREEN_REDIRECT_ALWAYS: The actor is always redirected through an FBO. CLUTTER_OFFSCREEN_REDIRECT_ONLY_FOR_OPACITY: The actor is only redirected through an FBO if the paint opacity is not 255. This value would be used if the actor wants correct opacity. It will avoid the overhead of using an FBO whenever the actor is fully opaque. The property is implemented by installing a ClutterFlattenEffect. ClutterFlattenEffect is a new internal class which subclasses ClutterOffscreen to redirect the painting to an FBO. When ClutterOffscreen paints, the effect sets an opacity override on the actor so that the image will always contain the actor at full opacity. The opacity is then applied to the resulting image before painting it to the stage. This means the actor does not need to be redrawn while the opacity is being animated. The effect has a high internal priority so that it will always occur before any other effects and it gets hidden from the application.
2011-02-28 15:05:51 +00:00
g_object_notify_by_pspec (G_OBJECT (self),
obj_props[PROP_OFFSCREEN_REDIRECT]);
}
}
/**
* clutter_actor_get_offscreen_redirect:
* @self: a #ClutterActor
*
* Retrieves whether to redirect the actor to an offscreen buffer, as
* set by clutter_actor_set_offscreen_redirect().
*
* Return value: the value of the offscreen-redirect property of the actor
*
* Since: 1.8
*/
ClutterOffscreenRedirect
clutter_actor_get_offscreen_redirect (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
return self->priv->offscreen_redirect;
}
/**
* clutter_actor_set_name:
* @self: A #ClutterActor
* @name: Textual tag to apply to actor
*
* Sets the given name to @self. The name can be used to identify
* a #ClutterActor.
*/
void
clutter_actor_set_name (ClutterActor *self,
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
const gchar *name)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
g_free (self->priv->name);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
self->priv->name = g_strdup (name);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_NAME]);
}
/**
* clutter_actor_get_name:
* @self: A #ClutterActor
*
* Retrieves the name of @self.
*
* Return value: the name of the actor, or %NULL. The returned string is
* owned by the actor and should not be modified or freed.
*/
const gchar *
clutter_actor_get_name (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
return self->priv->name;
}
/**
* clutter_actor_get_gid:
* @self: A #ClutterActor
*
2006-06-21 23:02:20 +00:00
* Retrieves the unique id for @self.
*
* Return value: Globally unique value for this object instance.
*
* Since: 0.6
*
* Deprecated: 1.8: The id is not used any longer.
*/
guint32
clutter_actor_get_gid (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
return self->priv->id;
}
/**
* clutter_actor_set_depth:
* @self: a #ClutterActor
2006-06-21 23:02:20 +00:00
* @depth: Z co-ord
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Sets the Z coordinate of @self to @depth.
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The unit used by @depth is dependant on the perspective setup. See
* also clutter_stage_set_perspective().
*/
void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_set_depth (ClutterActor *self,
gfloat depth)
{
2006-12-12 Emmanuele Bassi <ebassi@openedhand.com> Rework part of the show/hide machinery. Allow groups sub-classes and composite actors to override show_all/hide_all in order to decide which children they wish to show/hide. This means that if an actor overrides the default show/hide virtual methods, it'll have to chain up to the parent class show/hide. While we're at it, provide the fully recursive clutter_actor_show_all() and clutter_actor_hide_all() methods. * clutter/clutter-behaviour-path.c: Add apidoc for the ClutterKnot functions; add pathological equality case for clutter_knot_equal(). * clutter/clutter-event.h: * clutter/clutter-feature.h: * clutter/clutter-behaviour.c: * clutter/clutter-behaviour-scale.c:Fix parameters name so that gtk-doc doesn't complain. * clutter/clutter-actor.c: * clutter/clutter-event.c: Add apidoc * clutter/clutter-actor.h: * clutter/clutter-actor.c: Add a clutter_actor_show_all() and a clutter_actor_hide_all() functions; provide a mechanism for groups and composited actors to programmatically select what to show/hide when clutter_actor_show_all() and clutter_actor_hide_all() are called. If you are overriding the ClutterActor::show or the ClutterActor::hide virtual methods you should chain up with the parent class. * clutter/clutter-group.c: Override show_all and hide_all and recursively show/hide every child inside the group; clutter_group_show_all() and clutter_group_hide_all() remain as non recursive versions of clutter_actor_show_all() and clutter_actor_hide_all() (maybe we should rename them in order to avoid name clashes with the bindings). * clutter/clutter-stage.c: * clutter/clutter-texture.c: Chain up with parent class show and hide vfuncs. * clutter/clutter-clone-texture.h: * clutter/clutter-clone-texture.c: Provide API for changing the parent texture of a clone texture actor. * examples/behave.c: * examples/super-oh.c: * examples/test.c: Use clutter_actor_show_all() instead of clutter_group_show_all().
2006-12-12 20:20:04 +00:00
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
if (priv->z != depth)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
/* Sets Z value - XXX 2.0: should we invert? */
priv->z = depth;
priv->transform_valid = FALSE;
/* FIXME - remove this crap; sadly, there are still containers
* in Clutter that depend on this utter brain damage
*/
clutter_container_sort_depth_order (CLUTTER_CONTAINER (self));
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_queue_redraw (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_DEPTH]);
}
}
/**
* clutter_actor_get_depth:
* @self: a #ClutterActor
*
* Retrieves the depth of @self.
*
* Return value: the depth of the actor
*/
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat
clutter_actor_get_depth (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), -1);
return self->priv->z;
}
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_set_rotation:
* @self: a #ClutterActor
* @axis: the axis of rotation
* @angle: the angle of rotation
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* @x: X coordinate of the rotation center
* @y: Y coordinate of the rotation center
* @z: Z coordinate of the rotation center
*
* Sets the rotation angle of @self around the given axis.
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* The rotation center coordinates used depend on the value of @axis:
* <itemizedlist>
* <listitem><para>%CLUTTER_X_AXIS requires @y and @z</para></listitem>
* <listitem><para>%CLUTTER_Y_AXIS requires @x and @z</para></listitem>
* <listitem><para>%CLUTTER_Z_AXIS requires @x and @y</para></listitem>
* </itemizedlist>
*
* The rotation coordinates are relative to the anchor point of the
* actor, set using clutter_actor_set_anchor_point(). If no anchor
* point is set, the upper left corner is assumed as the origin.
*
* Since: 0.8
*/
void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_set_rotation (ClutterActor *self,
ClutterRotateAxis axis,
gdouble angle,
gfloat x,
gfloat y,
gfloat z)
{
ClutterVertex v;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
v.x = x;
v.y = y;
v.z = z;
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_set_rotation_angle_internal (self, axis, angle);
clutter_actor_set_rotation_center_internal (self, axis, &v);
g_object_thaw_notify (G_OBJECT (self));
}
/**
* clutter_actor_set_z_rotation_from_gravity:
* @self: a #ClutterActor
* @angle: the angle of rotation
* @gravity: the center point of the rotation
*
* Sets the rotation angle of @self around the Z axis using the center
* point specified as a compass point. For example to rotate such that
* the center of the actor remains static you can use
* %CLUTTER_GRAVITY_CENTER. If the actor changes size the center point
* will move accordingly.
*
* Since: 1.0
*/
void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_set_z_rotation_from_gravity (ClutterActor *self,
gdouble angle,
ClutterGravity gravity)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (gravity == CLUTTER_GRAVITY_NONE)
clutter_actor_set_rotation (self, CLUTTER_Z_AXIS, angle, 0, 0, 0);
else
{
GObject *obj = G_OBJECT (self);
ClutterTransformInfo *info;
info = _clutter_actor_get_transform_info (self);
g_object_freeze_notify (obj);
clutter_actor_set_rotation_angle_internal (self, CLUTTER_Z_AXIS, angle);
clutter_anchor_coord_set_gravity (&info->rz_center, gravity);
g_object_notify_by_pspec (obj, obj_props[PROP_ROTATION_CENTER_Z_GRAVITY]);
g_object_notify_by_pspec (obj, obj_props[PROP_ROTATION_CENTER_Z]);
g_object_thaw_notify (obj);
}
}
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_get_rotation:
* @self: a #ClutterActor
* @axis: the axis of rotation
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* @x: (out): return value for the X coordinate of the center of rotation
* @y: (out): return value for the Y coordinate of the center of rotation
* @z: (out): return value for the Z coordinate of the center of rotation
*
* Retrieves the angle and center of rotation on the given axis,
* set using clutter_actor_set_rotation().
*
* Return value: the angle of rotation
*
* Since: 0.8
*/
gdouble
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_get_rotation (ClutterActor *self,
ClutterRotateAxis axis,
gfloat *x,
gfloat *y,
gfloat *z)
{
const ClutterTransformInfo *info;
const AnchorCoord *anchor_coord;
gdouble retval = 0;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
info = _clutter_actor_get_transform_info_or_defaults (self);
switch (axis)
{
case CLUTTER_X_AXIS:
anchor_coord = &info->rx_center;
retval = info->rx_angle;
break;
case CLUTTER_Y_AXIS:
anchor_coord = &info->ry_center;
retval = info->ry_angle;
break;
case CLUTTER_Z_AXIS:
anchor_coord = &info->rz_center;
retval = info->rz_angle;
break;
default:
anchor_coord = NULL;
retval = 0.0;
break;
}
clutter_anchor_coord_get_units (self, anchor_coord, x, y, z);
return retval;
}
/**
* clutter_actor_get_z_rotation_gravity:
* @self: A #ClutterActor
*
* Retrieves the center for the rotation around the Z axis as a
* compass direction. If the center was specified in pixels or units
* this will return %CLUTTER_GRAVITY_NONE.
*
* Return value: the Z rotation center
*
* Since: 1.0
*/
ClutterGravity
clutter_actor_get_z_rotation_gravity (ClutterActor *self)
{
const ClutterTransformInfo *info;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0.0);
info = _clutter_actor_get_transform_info_or_defaults (self);
return clutter_anchor_coord_get_gravity (&info->rz_center);
}
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_set_clip:
* @self: A #ClutterActor
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* @xoff: X offset of the clip rectangle
* @yoff: Y offset of the clip rectangle
* @width: Width of the clip rectangle
* @height: Height of the clip rectangle
*
* Sets clip area for @self. The clip area is always computed from the
* upper left corner of the actor, even if the anchor point is set
* otherwise.
*
* Since: 0.6
*/
void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_set_clip (ClutterActor *self,
gfloat xoff,
gfloat yoff,
gfloat width,
gfloat height)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
if (priv->has_clip &&
priv->clip.x == xoff &&
priv->clip.y == yoff &&
priv->clip.width == width &&
priv->clip.height == height)
return;
priv->clip.x = xoff;
priv->clip.y = yoff;
priv->clip.width = width;
priv->clip.height = height;
priv->has_clip = TRUE;
clutter_actor_queue_redraw (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_HAS_CLIP]);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_CLIP]);
}
/**
* clutter_actor_remove_clip:
* @self: A #ClutterActor
*
* Removes clip area from @self.
*/
void
clutter_actor_remove_clip (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (!self->priv->has_clip)
return;
self->priv->has_clip = FALSE;
clutter_actor_queue_redraw (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_HAS_CLIP]);
}
/**
* clutter_actor_has_clip:
* @self: a #ClutterActor
*
* Determines whether the actor has a clip area set or not.
*
* Return value: %TRUE if the actor has a clip area set.
*
* Since: 0.1.1
*/
gboolean
clutter_actor_has_clip (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
return self->priv->has_clip;
}
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_get_clip:
* @self: a #ClutterActor
2009-07-02 13:21:16 +00:00
* @xoff: (out) (allow-none): return location for the X offset of
* the clip rectangle, or %NULL
* @yoff: (out) (allow-none): return location for the Y offset of
* the clip rectangle, or %NULL
* @width: (out) (allow-none): return location for the width of
* the clip rectangle, or %NULL
* @height: (out) (allow-none): return location for the height of
* the clip rectangle, or %NULL
*
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* Gets the clip area for @self, if any is set
*
* Since: 0.6
*/
void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_get_clip (ClutterActor *self,
gfloat *xoff,
gfloat *yoff,
gfloat *width,
gfloat *height)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
if (!priv->has_clip)
return;
if (xoff != NULL)
*xoff = priv->clip.x;
if (yoff != NULL)
*yoff = priv->clip.y;
if (width != NULL)
*width = priv->clip.width;
if (height != NULL)
*height = priv->clip.height;
}
/**
* clutter_actor_get_children:
* @self: a #ClutterActor
*
* Retrieves the list of children of @self.
*
* Return value: (transfer container) (element-type ClutterActor): A newly
* allocated #GList of #ClutterActor<!-- -->s. Use g_list_free() when
* done.
*
* Since: 1.10
*/
GList *
clutter_actor_get_children (ClutterActor *self)
{
ClutterActor *iter;
GList *res;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
/* we walk the list backward so that we can use prepend(),
* which is O(1)
*/
for (iter = self->priv->last_child, res = NULL;
iter != NULL;
iter = iter->priv->prev_sibling)
{
res = g_list_prepend (res, iter);
}
return res;
}
/*< private >
* insert_child_at_depth:
* @self: a #ClutterActor
* @child: a #ClutterActor
*
* Inserts @child inside the list of children held by @self, using
* the depth as the insertion criteria.
*
* This sadly makes the insertion not O(1), but we can keep the
* list sorted so that the painters algorithm we use for painting
* the children will work correctly.
*/
static void
insert_child_at_depth (ClutterActor *self,
ClutterActor *child,
gpointer dummy G_GNUC_UNUSED)
{
ClutterActor *iter;
child->priv->parent = self;
/* special-case the first child */
if (self->priv->n_children == 0)
{
self->priv->first_child = child;
self->priv->last_child = child;
child->priv->next_sibling = NULL;
child->priv->prev_sibling = NULL;
return;
}
/* Find the right place to insert the child so that it will still be
sorted and the child will be after all of the actors at the same
dept */
for (iter = self->priv->first_child;
iter != NULL;
iter = iter->priv->next_sibling)
{
if (iter->priv->z > child->priv->z)
break;
}
if (iter != NULL)
{
ClutterActor *tmp = iter->priv->prev_sibling;
if (tmp != NULL)
tmp->priv->next_sibling = child;
/* Insert the node before the found one */
child->priv->prev_sibling = iter->priv->prev_sibling;
child->priv->next_sibling = iter;
iter->priv->prev_sibling = child;
}
else
{
ClutterActor *tmp = self->priv->last_child;
if (tmp != NULL)
tmp->priv->next_sibling = child;
/* insert the node at the end of the list */
child->priv->prev_sibling = self->priv->last_child;
child->priv->next_sibling = NULL;
}
if (child->priv->prev_sibling == NULL)
self->priv->first_child = child;
if (child->priv->next_sibling == NULL)
self->priv->last_child = child;
}
static void
insert_child_at_index (ClutterActor *self,
ClutterActor *child,
gpointer data_)
{
gint index_ = GPOINTER_TO_INT (data_);
child->priv->parent = self;
if (index_ == 0)
{
ClutterActor *tmp = self->priv->first_child;
if (tmp != NULL)
tmp->priv->prev_sibling = child;
child->priv->prev_sibling = NULL;
child->priv->next_sibling = tmp;
}
else if (index_ < 0)
{
ClutterActor *tmp = self->priv->last_child;
if (tmp != NULL)
tmp->priv->next_sibling = child;
child->priv->prev_sibling = tmp;
child->priv->next_sibling = NULL;
}
else
{
ClutterActor *iter;
int i;
for (iter = self->priv->first_child, i = 0;
iter != NULL;
iter = iter->priv->next_sibling, i += 1)
{
if (index_ == i)
{
ClutterActor *tmp = iter->priv->prev_sibling;
child->priv->prev_sibling = tmp;
child->priv->next_sibling = iter;
iter->priv->prev_sibling = child;
if (tmp != NULL)
tmp->priv->next_sibling = child;
break;
}
}
}
if (child->priv->prev_sibling == NULL)
self->priv->first_child = child;
if (child->priv->next_sibling == NULL)
self->priv->last_child = child;
}
static void
insert_child_above (ClutterActor *self,
ClutterActor *child,
gpointer data)
{
ClutterActor *sibling = data;
child->priv->parent = self;
if (sibling == NULL)
sibling = self->priv->last_child;
child->priv->prev_sibling = sibling;
if (sibling != NULL)
{
ClutterActor *tmp = sibling->priv->next_sibling;
child->priv->next_sibling = tmp;
if (tmp != NULL)
tmp->priv->prev_sibling = child;
sibling->priv->next_sibling = child;
}
else
child->priv->next_sibling = NULL;
if (child->priv->prev_sibling == NULL)
self->priv->first_child = child;
if (child->priv->next_sibling == NULL)
self->priv->last_child = child;
}
static void
insert_child_below (ClutterActor *self,
ClutterActor *child,
gpointer data)
{
ClutterActor *sibling = data;
child->priv->parent = self;
if (sibling == NULL)
sibling = self->priv->first_child;
child->priv->next_sibling = sibling;
if (sibling != NULL)
{
ClutterActor *tmp = sibling->priv->prev_sibling;
child->priv->prev_sibling = tmp;
if (tmp != NULL)
tmp->priv->next_sibling = child;
sibling->priv->prev_sibling = child;
}
else
child->priv->prev_sibling = NULL;
if (child->priv->prev_sibling == NULL)
self->priv->first_child = child;
if (child->priv->next_sibling == NULL)
self->priv->last_child = child;
}
typedef void (* ClutterActorAddChildFunc) (ClutterActor *parent,
ClutterActor *child,
gpointer data);
typedef enum {
ADD_CHILD_CREATE_META = 1 << 0,
ADD_CHILD_EMIT_PARENT_SET = 1 << 1,
ADD_CHILD_EMIT_ACTOR_ADDED = 1 << 2,
ADD_CHILD_CHECK_STATE = 1 << 3,
ADD_CHILD_NOTIFY_FIRST_LAST = 1 << 4,
/* default flags for public API */
ADD_CHILD_DEFAULT_FLAGS = ADD_CHILD_CREATE_META |
ADD_CHILD_EMIT_PARENT_SET |
ADD_CHILD_EMIT_ACTOR_ADDED |
ADD_CHILD_CHECK_STATE |
ADD_CHILD_NOTIFY_FIRST_LAST,
/* flags for legacy/deprecated API */
ADD_CHILD_LEGACY_FLAGS = ADD_CHILD_EMIT_PARENT_SET |
ADD_CHILD_CHECK_STATE |
ADD_CHILD_NOTIFY_FIRST_LAST
} ClutterActorAddChildFlags;
/*< private >
* clutter_actor_add_child_internal:
* @self: a #ClutterActor
* @child: a #ClutterActor
* @flags: control flags for actions
* @add_func: delegate function
* @data: (closure): data to pass to @add_func
*
* Adds @child to the list of children of @self.
*
* The actual insertion inside the list is delegated to @add_func: this
* function will just set up the state, perform basic checks, and emit
* signals.
*
* The @flags argument is used to perform additional operations.
*/
static inline void
clutter_actor_add_child_internal (ClutterActor *self,
ClutterActor *child,
ClutterActorAddChildFlags flags,
ClutterActorAddChildFunc add_func,
gpointer data)
{
ClutterTextDirection text_dir;
gboolean create_meta;
gboolean emit_parent_set, emit_actor_added;
gboolean check_state;
gboolean notify_first_last;
ClutterActor *old_first_child, *old_last_child;
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (child->priv->parent != NULL)
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
{
g_warning ("Cannot set a parent on an actor which has a parent. "
"You must use clutter_actor_remove_child() first.");
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
return;
}
if (CLUTTER_ACTOR_IS_TOPLEVEL (child))
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
{
g_warning ("Cannot set a parent on a toplevel actor\n");
return;
}
if (CLUTTER_ACTOR_IN_DESTRUCTION (child))
{
g_warning ("Cannot set a parent currently being destroyed");
return;
}
create_meta = (flags & ADD_CHILD_CREATE_META) != 0;
emit_parent_set = (flags & ADD_CHILD_EMIT_PARENT_SET) != 0;
emit_actor_added = (flags & ADD_CHILD_EMIT_ACTOR_ADDED) != 0;
check_state = (flags & ADD_CHILD_CHECK_STATE) != 0;
notify_first_last = (flags & ADD_CHILD_NOTIFY_FIRST_LAST) != 0;
old_first_child = self->priv->first_child;
old_last_child = self->priv->last_child;
if (create_meta)
clutter_container_create_child_meta (CLUTTER_CONTAINER (self), child);
g_object_ref_sink (child);
child->priv->parent = NULL;
child->priv->next_sibling = NULL;
child->priv->prev_sibling = NULL;
/* delegate the actual insertion */
add_func (self, child, data);
g_assert (child->priv->parent == self);
self->priv->n_children += 1;
self->priv->age += 1;
/* if push_internal() has been called then we automatically set
* the flag on the actor
*/
if (self->priv->internal_child)
CLUTTER_SET_PRIVATE_FLAGS (child, CLUTTER_INTERNAL_CHILD);
/* clutter_actor_reparent() will emit ::parent-set for us */
if (emit_parent_set && !CLUTTER_ACTOR_IN_REPARENT (child))
g_signal_emit (child, actor_signals[PARENT_SET], 0, NULL);
if (check_state)
{
/* If parent is mapped or realized, we need to also be mapped or
* realized once we're inside the parent.
*/
clutter_actor_update_map_state (child, MAP_STATE_CHECK);
/* propagate the parent's text direction to the child */
text_dir = clutter_actor_get_text_direction (self);
clutter_actor_set_text_direction (child, text_dir);
}
if (child->priv->show_on_set_parent)
clutter_actor_show (child);
if (CLUTTER_ACTOR_IS_MAPPED (child))
clutter_actor_queue_redraw (child);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
/* maintain the invariant that if an actor needs layout,
* its parents do as well
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*/
if (child->priv->needs_width_request ||
child->priv->needs_height_request ||
child->priv->needs_allocation)
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
{
/* we work around the short-circuiting we do
* in clutter_actor_queue_relayout() since we
* want to force a relayout
*/
child->priv->needs_width_request = TRUE;
child->priv->needs_height_request = TRUE;
child->priv->needs_allocation = TRUE;
clutter_actor_queue_relayout (child->priv->parent);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
}
if (emit_actor_added)
g_signal_emit_by_name (self, "actor-added", child);
if (notify_first_last)
{
if (old_first_child != self->priv->first_child)
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_FIRST_CHILD]);
if (old_last_child != self->priv->last_child)
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_LAST_CHILD]);
}
}
/**
* clutter_actor_add_child:
* @self: a #ClutterActor
* @child: a #ClutterActor
*
* Adds @child to the children of @self.
*
* This function will acquire a reference on @child that will only
* be released when calling clutter_actor_remove_child().
*
* This function will take into consideration the #ClutterActor:depth
* of @child, and will keep the list of children sorted.
*
* This function will emit the #ClutterContainer::actor-added signal
* on @self.
*
* Since: 1.10
*/
void
clutter_actor_add_child (ClutterActor *self,
ClutterActor *child)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (child));
g_return_if_fail (self != child);
g_return_if_fail (child->priv->parent == NULL);
clutter_actor_add_child_internal (self, child,
ADD_CHILD_DEFAULT_FLAGS,
insert_child_at_depth,
NULL);
}
/**
* clutter_actor_insert_child_at_index:
* @self: a #ClutterActor
* @child: a #ClutterActor
* @index_: the index
*
* Inserts @child into the list of children of @self, using the
* given @index_.
*
* This function will acquire a reference on @child that will only
* be released when calling clutter_actor_remove_child().
*
* This function will not take into consideration the #ClutterActor:depth
* of @child.
*
* This function will emit the #ClutterContainer::actor-added signal
* on @self.
*
* Since: 1.10
*/
void
clutter_actor_insert_child_at_index (ClutterActor *self,
ClutterActor *child,
gint index_)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (child));
g_return_if_fail (self != child);
g_return_if_fail (child->priv->parent == NULL);
clutter_actor_add_child_internal (self, child,
ADD_CHILD_DEFAULT_FLAGS,
insert_child_at_index,
GINT_TO_POINTER (index_));
}
/**
* clutter_actor_insert_child_above:
* @self: a #ClutterActor
* @child: a #ClutterActor
* @sibling: (allow-none): a child of @self, or %NULL
*
* Inserts @child into the list of children of @self, above another
* child of @self or, if @sibling is %NULL, above all the children
* of @self.
*
* This function will acquire a reference on @child that will only
* be released when calling clutter_actor_remove_child().
*
* This function will not take into consideration the #ClutterActor:depth
* of @child.
*
* This function will emit the #ClutterContainer::actor-added signal
* on @self.
*
* Since: 1.10
*/
void
clutter_actor_insert_child_above (ClutterActor *self,
ClutterActor *child,
ClutterActor *sibling)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (child));
g_return_if_fail (self != child);
g_return_if_fail (child != sibling);
g_return_if_fail (child->priv->parent == NULL);
g_return_if_fail (sibling == NULL ||
(CLUTTER_IS_ACTOR (sibling) &&
sibling->priv->parent == self));
clutter_actor_add_child_internal (self, child,
ADD_CHILD_DEFAULT_FLAGS,
insert_child_above,
sibling);
}
/**
* clutter_actor_insert_child_below:
* @self: a #ClutterActor
* @child: a #ClutterActor
* @sibling: (allow-none): a child of @self, or %NULL
*
* Inserts @child into the list of children of @self, below another
* child of @self or, if @sibling is %NULL, below all the children
* of @self.
*
* This function will acquire a reference on @child that will only
* be released when calling clutter_actor_remove_child().
*
* This function will not take into consideration the #ClutterActor:depth
* of @child.
*
* This function will emit the #ClutterContainer::actor-added signal
* on @self.
*
* Since: 1.10
*/
void
clutter_actor_insert_child_below (ClutterActor *self,
ClutterActor *child,
ClutterActor *sibling)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (child));
g_return_if_fail (self != child);
g_return_if_fail (child != sibling);
g_return_if_fail (child->priv->parent == NULL);
g_return_if_fail (sibling == NULL ||
(CLUTTER_IS_ACTOR (sibling) &&
sibling->priv->parent == self));
clutter_actor_add_child_internal (self, child,
ADD_CHILD_DEFAULT_FLAGS,
insert_child_below,
sibling);
}
/**
* clutter_actor_set_parent:
* @self: A #ClutterActor
* @parent: A new #ClutterActor parent
*
* Sets the parent of @self to @parent.
*
* This function will result in @parent acquiring a reference on @self,
* eventually by sinking its floating reference first. The reference
* will be released by clutter_actor_unparent().
*
* This function should only be called by legacy #ClutterActor<!-- -->s
* implementing the #ClutterContainer interface.
*
* Deprecated: 1.10: Use clutter_actor_add_child() instead.
*/
void
clutter_actor_set_parent (ClutterActor *self,
ClutterActor *parent)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (parent));
g_return_if_fail (self != parent);
g_return_if_fail (self->priv->parent == NULL);
/* as this function will be called inside ClutterContainer::add
* implementations or when building up a composite actor, we have
* to preserve the old behaviour, and not create child meta or
* emit the ::actor-added signal, to avoid recursion or double
* emissions
*/
clutter_actor_add_child_internal (parent, self,
ADD_CHILD_LEGACY_FLAGS,
insert_child_at_depth,
NULL);
}
/**
* clutter_actor_get_parent:
* @self: A #ClutterActor
*
2006-06-21 23:02:20 +00:00
* Retrieves the parent of @self.
*
2009-07-02 13:21:16 +00:00
* Return Value: (transfer none): The #ClutterActor parent, or %NULL
* if no parent is set
*/
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
ClutterActor *
clutter_actor_get_parent (ClutterActor *self)
{
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
return self->priv->parent;
}
/**
* clutter_actor_get_paint_visibility:
* @self: A #ClutterActor
*
* Retrieves the 'paint' visibility of an actor recursively checking for non
* visible parents.
*
* This is by definition the same as %CLUTTER_ACTOR_IS_MAPPED.
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
*
* Return Value: %TRUE if the actor is visibile and will be painted.
*
* Since: 0.8.4
*/
gboolean
clutter_actor_get_paint_visibility (ClutterActor *actor)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (actor), FALSE);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
return CLUTTER_ACTOR_IS_MAPPED (actor);
}
/**
* clutter_actor_remove_child:
* @self: a #ClutterActor
* @child: a #ClutterActor
*
* Removes @child from the children of @self.
*
* This function will release the reference added by
* clutter_actor_add_child(), so if you want to keep using @child
* you will have to acquire a referenced on it before calling this
* function.
*
* This function will emit the #ClutterContainer::actor-removed
* signal on @self.
*
* Since: 1.10
*/
void
clutter_actor_remove_child (ClutterActor *self,
ClutterActor *child)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (child));
g_return_if_fail (self != child);
g_return_if_fail (child->priv->parent != NULL);
g_return_if_fail (child->priv->parent == self);
clutter_actor_remove_child_internal (self, child,
REMOVE_CHILD_DEFAULT_FLAGS);
}
/**
* clutter_actor_remove_all_children:
* @self: a #ClutterActor
*
* Removes all children of @self.
*
* This function releases the reference added by inserting a child actor
* in the list of children of @self.
*
* If the reference count of a child drops to zero, the child will be
* destroyed. If you want to ensure the destruction of all the children
* of @self, use clutter_actor_destroy_all_children().
*
* Since: 1.10
*/
void
clutter_actor_remove_all_children (ClutterActor *self)
{
ClutterActorIter iter;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (self->priv->n_children == 0)
return;
clutter_actor_iter_init (&iter, self);
while (clutter_actor_iter_next (&iter, NULL))
clutter_actor_iter_remove (&iter);
/* sanity check */
g_assert (self->priv->first_child == NULL);
g_assert (self->priv->last_child == NULL);
g_assert (self->priv->n_children == 0);
}
/**
* clutter_actor_destroy_all_children:
* @self: a #ClutterActor
*
* Destroys all children of @self.
*
* This function releases the reference added by inserting a child
* actor in the list of children of @self, and ensures that the
* #ClutterActor::destroy signal is emitted on each child of the
* actor.
*
* By default, #ClutterActor will emit the #ClutterActor::destroy signal
* when its reference count drops to 0; the default handler of the
* #ClutterActor::destroy signal will destroy all the children of an
* actor. This function ensures that all children are destroyed, instead
* of just removed from @self, unlike clutter_actor_remove_all_children()
* which will merely release the reference and remove each child.
*
* Unless you acquired an additional reference on each child of @self
* prior to calling clutter_actor_remove_all_children() and want to reuse
* the actors, you should use clutter_actor_destroy_all_children() in
* order to make sure that children are destroyed and signal handlers
* are disconnected even in cases where circular references prevent this
* from automatically happening through reference counting alone.
*
* Since: 1.10
*/
void
clutter_actor_destroy_all_children (ClutterActor *self)
{
ClutterActorIter iter;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (self->priv->n_children == 0)
return;
clutter_actor_iter_init (&iter, self);
while (clutter_actor_iter_next (&iter, NULL))
clutter_actor_iter_destroy (&iter);
/* sanity check */
g_assert (self->priv->first_child == NULL);
g_assert (self->priv->last_child == NULL);
g_assert (self->priv->n_children == 0);
}
typedef struct _InsertBetweenData {
ClutterActor *prev_sibling;
ClutterActor *next_sibling;
} InsertBetweenData;
static void
insert_child_between (ClutterActor *self,
ClutterActor *child,
gpointer data_)
{
InsertBetweenData *data = data_;
ClutterActor *prev_sibling = data->prev_sibling;
ClutterActor *next_sibling = data->next_sibling;
child->priv->parent = self;
child->priv->prev_sibling = prev_sibling;
child->priv->next_sibling = next_sibling;
if (prev_sibling != NULL)
prev_sibling->priv->next_sibling = child;
if (next_sibling != NULL)
next_sibling->priv->prev_sibling = child;
if (child->priv->prev_sibling == NULL)
self->priv->first_child = child;
if (child->priv->next_sibling == NULL)
self->priv->last_child = child;
}
/**
* clutter_actor_replace_child:
* @self: a #ClutterActor
* @old_child: the child of @self to replace
* @new_child: the #ClutterActor to replace @old_child
*
* Replaces @old_child with @new_child in the list of children of @self.
*
* Since: 1.10
*/
void
clutter_actor_replace_child (ClutterActor *self,
ClutterActor *old_child,
ClutterActor *new_child)
{
ClutterActor *prev_sibling, *next_sibling;
InsertBetweenData clos;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (old_child));
g_return_if_fail (old_child->priv->parent == self);
g_return_if_fail (CLUTTER_IS_ACTOR (new_child));
g_return_if_fail (old_child != new_child);
g_return_if_fail (new_child != self);
g_return_if_fail (new_child->priv->parent == NULL);
prev_sibling = old_child->priv->prev_sibling;
next_sibling = old_child->priv->next_sibling;
clutter_actor_remove_child_internal (self, old_child,
REMOVE_CHILD_DEFAULT_FLAGS);
clos.prev_sibling = prev_sibling;
clos.next_sibling = next_sibling;
clutter_actor_add_child_internal (self, new_child,
ADD_CHILD_DEFAULT_FLAGS,
insert_child_between,
&clos);
}
/**
* clutter_actor_unparent:
* @self: a #ClutterActor
*
* Removes the parent of @self.
*
* This will cause the parent of @self to release the reference
* acquired when calling clutter_actor_set_parent(), so if you
* want to keep @self you will have to acquire a reference of
* your own, through g_object_ref().
*
* This function should only be called by legacy #ClutterActor<!-- -->s
* implementing the #ClutterContainer interface.
*
* Since: 0.1.1
*
* Deprecated: 1.10: Use clutter_actor_remove_child() instead.
*/
void
clutter_actor_unparent (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (self->priv->parent == NULL)
return;
clutter_actor_remove_child_internal (self->priv->parent, self,
REMOVE_CHILD_LEGACY_FLAGS);
2006-07-06 Emmanuele Bassi <ebassi@openedhand.com> Big rework of the actor management semantics: now ClutterActor objects behave like GtkObjects - that is they have an initial "floating" reference that gets "sunk" when they are added to a ClutterGroup. This makes a group responsible of de-allocating each actor inside it, so you just have to destroy the group to get every child actor destroyed. Also, now you can do: clutter_group_add (group, clutter_video_texture_new ()); without having to care about reference counting and explicit unreffing. * clutter/clutter-private.h: Add private flags setter and getter macros. * clutter/clutter-actor.h: * clutter/clutter-actor.c: Clean up; inherit from GInitiallyUnowned; add a "visible" property; add the "destroy", "show" and "hide" signals to ClutterActorClass. (clutter_actor_show), (clutter_actor_hide): Refactor a bit; emit the "show" and "hide" signals. (clutter_actor_set_property), (clutter_actor_get_property), (clutter_actor_class_init): Implement the "visible" property; add signals. (clutter_actor_finalize): Do not leak the actor's name, if it is set. (clutter_actor_dispose): Emit the "destroy" signal here. (clutter_actor_init): Sink the initial floating flag if needed. (clutter_actor_destroy): Add a function to explicitely destroy a ClutterActor. (clutter_actor_set_parent), (clutter_actor_get_parent), (clutter_actor_unparent): Make set_parent require a valid parent; add unparent; check on get_parent; ref_sink the actor when setting its parent and unref it when unsetting it. Probably we'll need a function that does reparenting as unparent+set_parent in a single shot. * clutter/clutter-group.h: * clutter/clutter-group.c (clutter_group_dispose), (clutter_group_finalize), (clutter_group_add), (clutter_group_remove): Make the group destroy its children when disposing it; clean up, and use the newly-available clutter_actor_unparent(). * clutter/clutter-stage.h: * clutter/clutter-stage.c (clutter_stage_init): ClutterStage is a top-level actor; clean up. * clutter/clutter-video-texture.h: * clutter/clutter-video-texture.c: Clean up. * examples/super-oh.c: * examples/test.c: * examples/video-player.c: * examples/test-text.c: * examples/video-cube.c: Remove the g_object_unref() call, as the ClutterStage object is destroyed on clutter_main_quit().
2006-07-06 17:52:57 +00:00
}
/**
* clutter_actor_reparent:
* @self: a #ClutterActor
* @new_parent: the new #ClutterActor parent
*
* Resets the parent actor of @self.
*
* This function is logically equivalent to calling clutter_actor_unparent()
* and clutter_actor_set_parent(), but more efficiently implemented, as it
* ensures the child is not finalized when unparented, and emits the
* #ClutterActor::parent-set signal only once.
*
* In reality, calling this function is less useful than it sounds, as some
* application code may rely on changes in the intermediate state between
* removal and addition of the actor from its old parent to the @new_parent.
* Thus, it is strongly encouraged to avoid using this function in application
* code.
*
* Since: 0.2
*
* Deprecated: 1.10: Use clutter_actor_remove_child() and
* clutter_actor_add_child() instead; remember to take a reference on
* the actor being removed before calling clutter_actor_remove_child()
* to avoid the reference count dropping to zero and the actor being
* destroyed.
*/
void
clutter_actor_reparent (ClutterActor *self,
ClutterActor *new_parent)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (new_parent));
g_return_if_fail (self != new_parent);
if (CLUTTER_ACTOR_IS_TOPLEVEL (self))
{
g_warning ("Cannot set a parent on a toplevel actor");
return;
}
if (CLUTTER_ACTOR_IN_DESTRUCTION (self))
{
g_warning ("Cannot set a parent currently being destroyed");
return;
}
priv = self->priv;
if (priv->parent != new_parent)
{
ClutterActor *old_parent;
CLUTTER_SET_PRIVATE_FLAGS (self, CLUTTER_IN_REPARENT);
old_parent = priv->parent;
g_object_ref (self);
if (old_parent != NULL)
{
/* go through the Container implementation if this is a regular
* child and not an internal one
*/
if (!CLUTTER_ACTOR_IS_INTERNAL_CHILD (self))
{
ClutterContainer *parent = CLUTTER_CONTAINER (old_parent);
/* this will have to call unparent() */
clutter_container_remove_actor (parent, self);
}
else
clutter_actor_remove_child_internal (old_parent, self,
REMOVE_CHILD_LEGACY_FLAGS);
}
/* Note, will call set_parent() */
if (!CLUTTER_ACTOR_IS_INTERNAL_CHILD (self))
clutter_container_add_actor (CLUTTER_CONTAINER (new_parent), self);
else
clutter_actor_add_child_internal (new_parent, self,
ADD_CHILD_LEGACY_FLAGS,
insert_child_at_depth,
NULL);
/* we emit the ::parent-set signal once */
g_signal_emit (self, actor_signals[PARENT_SET], 0, old_parent);
CLUTTER_UNSET_PRIVATE_FLAGS (self, CLUTTER_IN_REPARENT);
Enforce invariants on mapped, realized, visibility states Bug 1138 - No trackable "mapped" state * Add a VISIBLE flag tracking application programmer's expected showing-state for the actor, allowing us to always ensure we keep what the app wants while tracking internal implementation state separately. * Make MAPPED reflect whether the actor will be painted; add notification on a ClutterActor::mapped property. Keep MAPPED state updated as the actor is shown, ancestors are shown, actor is reparented, etc. * Require a stage and realized parents to realize; this means at realization time the correct window system and GL resources are known. But unparented actors can no longer be realized. * Allow children to be unrealized even if parent is realized. Otherwise in effect either all actors or no actors are realized, i.e. it becomes a stage-global flag. * Allow clutter_actor_realize() to "fail" if not inside a toplevel * Rework clutter_actor_unrealize() so internally we have a flavor that does not mess with visibility flag * Add _clutter_actor_rerealize() to encapsulate a somewhat tricky operation we were doing in a couple of places * Do not realize/unrealize children in ClutterGroup, ClutterActor already does it * Do not realize impl by hand in clutter_stage_show(), since showing impl already does that * Do not unrealize in various dispose() methods, since ClutterActor dispose implementation already does it and chaining up is mandatory * ClutterTexture uses COGL while unrealizable (before it's added to a stage). Previously this breakage was affecting ClutterActor because we had to allow realize outside a stage. Move the breakage to ClutterTexture, by making ClutterTexture just use COGL while not realized. * Unrealize before we set parent to NULL in clutter_actor_unparent(). This means unrealize() implementations can get to the stage. Because actors need the stage in order to detach from stage. * Update clutter-actor-invariants.txt to reflect latest changes * Remove explicit hide/unrealize from ClutterActor::dispose since unparent already forces those Instead just assert that unparent() occurred and did the right thing. * Check whether parent implements unrealize before chaining up Needed because ClutterGroup no longer has to implement unrealize. * Perform unrealize in the default handler for the signal. This allows non-containers that have children to work properly, and allows containers to override how it's done. * Add map/unmap virtual methods and set MAPPED flag on self and children in there. This allows subclasses to hook map/unmap. These are not signals, because notify::mapped is better for anything it's legitimate for a non-subclass to do. Signed-off-by: Emmanuele Bassi <ebassi@linux.intel.com>
2009-04-02 13:16:43 +00:00
/* the IN_REPARENT flag suspends state updates */
clutter_actor_update_map_state (self, MAP_STATE_CHECK);
g_object_unref (self);
}
}
/**
* clutter_actor_contains:
* @self: A #ClutterActor
* @descendant: A #ClutterActor, possibly contained in @self
*
* Determines if @descendant is contained inside @self (either as an
* immediate child, or as a deeper descendant). If @self and
* @descendant point to the same actor then it will also return %TRUE.
*
* Return value: whether @descendent is contained within @self
*
* Since: 1.4
*/
gboolean
clutter_actor_contains (ClutterActor *self,
ClutterActor *descendant)
{
ClutterActor *actor;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
g_return_val_if_fail (CLUTTER_IS_ACTOR (descendant), FALSE);
for (actor = descendant; actor; actor = actor->priv->parent)
if (actor == self)
return TRUE;
return FALSE;
}
/**
2012-01-03 14:31:06 +00:00
* clutter_actor_set_child_above_sibling:
* @self: a #ClutterActor
* @child: a #ClutterActor child of @self
* @sibling: (allow-none): a #ClutterActor child of @self, or %NULL
*
* Sets @child to be above @sibling in the list of children of @self.
*
* If @sibling is %NULL, @child will be the new last child of @self.
*
* This function is logically equivalent to removing @child and using
* clutter_actor_insert_child_above(), but it will not emit signals
* or change state on @child.
*
* Since: 1.10
*/
void
clutter_actor_set_child_above_sibling (ClutterActor *self,
ClutterActor *child,
ClutterActor *sibling)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (child));
g_return_if_fail (child->priv->parent == self);
g_return_if_fail (child != sibling);
g_return_if_fail (sibling == NULL || CLUTTER_IS_ACTOR (sibling));
if (sibling != NULL)
g_return_if_fail (sibling->priv->parent == self);
/* we don't want to change the state of child, or emit signals, or
* regenerate ChildMeta instances here, but we still want to follow
* the correct sequence of steps encoded in remove_child() and
* add_child(), so that correctness is ensured, and we only go
* through one known code path.
*/
g_object_ref (child);
clutter_actor_remove_child_internal (self, child, 0);
clutter_actor_add_child_internal (self, child,
ADD_CHILD_NOTIFY_FIRST_LAST,
insert_child_above,
sibling);
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_set_child_below_sibling:
* @self: a #ClutterActor
* @child: a #ClutterActor child of @self
* @sibling: (allow-none): a #ClutterActor child of @self, or %NULL
*
* Sets @child to be below @sibling in the list of children of @self.
*
* If @sibling is %NULL, @child will be the new first child of @self.
*
* This function is logically equivalent to removing @self and using
* clutter_actor_insert_child_below(), but it will not emit signals
* or change state on @child.
*
* Since: 1.10
*/
void
clutter_actor_set_child_below_sibling (ClutterActor *self,
ClutterActor *child,
ClutterActor *sibling)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (child));
g_return_if_fail (child->priv->parent == self);
g_return_if_fail (child != sibling);
g_return_if_fail (sibling == NULL || CLUTTER_IS_ACTOR (sibling));
if (sibling != NULL)
g_return_if_fail (sibling->priv->parent == self);
/* see the comment in set_child_above_sibling() */
g_object_ref (child);
clutter_actor_remove_child_internal (self, child, 0);
clutter_actor_add_child_internal (self, child,
ADD_CHILD_NOTIFY_FIRST_LAST,
insert_child_below,
sibling);
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_set_child_at_index:
* @self: a #ClutterActor
* @child: a #ClutterActor child of @self
* @index_: the new index for @child
*
* Changes the index of @child in the list of children of @self.
*
* This function is logically equivalent to removing @child and
* calling clutter_actor_insert_child_at_index(), but it will not
* emit signals or change state on @child.
*
* Since: 1.10
*/
void
clutter_actor_set_child_at_index (ClutterActor *self,
ClutterActor *child,
gint index_)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTOR (child));
g_return_if_fail (child->priv->parent == self);
g_return_if_fail (index_ <= self->priv->n_children);
g_object_ref (child);
clutter_actor_remove_child_internal (self, child, 0);
clutter_actor_add_child_internal (self, child,
ADD_CHILD_NOTIFY_FIRST_LAST,
insert_child_at_index,
GINT_TO_POINTER (index_));
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_raise:
* @self: A #ClutterActor
* @below: (allow-none): A #ClutterActor to raise above.
*
2006-06-21 23:02:20 +00:00
* Puts @self above @below.
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Both actors must have the same parent, and the parent must implement
* the #ClutterContainer interface
*
* This function calls clutter_container_raise_child() internally.
*
* Deprecated: 1.10: Use clutter_actor_set_child_above_sibling() instead.
*/
void
clutter_actor_raise (ClutterActor *self,
ClutterActor *below)
{
ClutterActor *parent;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
parent = clutter_actor_get_parent (self);
if (parent == NULL)
{
g_warning ("%s: Actor '%s' is not inside a container",
G_STRFUNC,
_clutter_actor_get_debug_name (self));
return;
}
if (below != NULL)
{
if (parent != clutter_actor_get_parent (below))
{
g_warning ("%s Actor '%s' is not in the same container as "
"actor '%s'",
G_STRFUNC,
_clutter_actor_get_debug_name (self),
_clutter_actor_get_debug_name (below));
return;
}
}
clutter_container_raise_child (CLUTTER_CONTAINER (parent), self, below);
}
/**
* clutter_actor_lower:
* @self: A #ClutterActor
* @above: (allow-none): A #ClutterActor to lower below
*
2006-06-21 23:02:20 +00:00
* Puts @self below @above.
*
* Both actors must have the same parent, and the parent must implement
* the #ClutterContainer interface.
*
* This function calls clutter_container_lower_child() internally.
*
* Deprecated: 1.10: Use clutter_actor_set_child_below_sibling() instead.
*/
void
clutter_actor_lower (ClutterActor *self,
ClutterActor *above)
{
ClutterActor *parent;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
parent = clutter_actor_get_parent (self);
if (parent == NULL)
{
g_warning ("%s: Actor of type %s is not inside a container",
G_STRFUNC,
_clutter_actor_get_debug_name (self));
return;
}
if (above)
{
if (parent != clutter_actor_get_parent (above))
{
g_warning ("%s: Actor '%s' is not in the same container as "
"actor '%s'",
G_STRFUNC,
_clutter_actor_get_debug_name (self),
_clutter_actor_get_debug_name (above));
return;
}
}
clutter_container_lower_child (CLUTTER_CONTAINER (parent), self, above);
}
/**
* clutter_actor_raise_top:
* @self: A #ClutterActor
*
2006-06-21 23:02:20 +00:00
* Raises @self to the top.
*
* This function calls clutter_actor_raise() internally.
*
* Deprecated: 1.10: Use clutter_actor_set_child_above_sibling() with
* a %NULL sibling, instead.
*/
void
clutter_actor_raise_top (ClutterActor *self)
{
clutter_actor_raise (self, NULL);
}
/**
* clutter_actor_lower_bottom:
* @self: A #ClutterActor
*
* Lowers @self to the bottom.
*
* This function calls clutter_actor_lower() internally.
*
* Deprecated: 1.10: Use clutter_actor_set_child_below_sibling() with
* a %NULL sibling, instead.
*/
void
clutter_actor_lower_bottom (ClutterActor *self)
{
clutter_actor_lower (self, NULL);
}
/*
* Event handling
*/
/**
* clutter_actor_event:
* @actor: a #ClutterActor
* @event: a #ClutterEvent
* @capture: TRUE if event in in capture phase, FALSE otherwise.
*
* This function is used to emit an event on the main stage.
* You should rarely need to use this function, except for
* synthetising events.
*
* Return value: the return value from the signal emission: %TRUE
* if the actor handled the event, or %FALSE if the event was
* not handled
*
* Since: 0.6
*/
gboolean
clutter_actor_event (ClutterActor *actor,
ClutterEvent *event,
gboolean capture)
{
gboolean retval = FALSE;
gint signal_num = -1;
g_return_val_if_fail (CLUTTER_IS_ACTOR (actor), FALSE);
g_return_val_if_fail (event != NULL, FALSE);
g_object_ref (actor);
if (capture)
{
g_signal_emit (actor, actor_signals[CAPTURED_EVENT], 0,
event,
&retval);
goto out;
}
g_signal_emit (actor, actor_signals[EVENT], 0, event, &retval);
if (!retval)
{
switch (event->type)
{
case CLUTTER_NOTHING:
break;
case CLUTTER_BUTTON_PRESS:
signal_num = BUTTON_PRESS_EVENT;
break;
case CLUTTER_BUTTON_RELEASE:
signal_num = BUTTON_RELEASE_EVENT;
break;
case CLUTTER_SCROLL:
signal_num = SCROLL_EVENT;
break;
case CLUTTER_KEY_PRESS:
signal_num = KEY_PRESS_EVENT;
break;
case CLUTTER_KEY_RELEASE:
signal_num = KEY_RELEASE_EVENT;
break;
case CLUTTER_MOTION:
signal_num = MOTION_EVENT;
break;
case CLUTTER_ENTER:
signal_num = ENTER_EVENT;
break;
case CLUTTER_LEAVE:
signal_num = LEAVE_EVENT;
break;
case CLUTTER_DELETE:
case CLUTTER_DESTROY_NOTIFY:
case CLUTTER_CLIENT_MESSAGE:
default:
signal_num = -1;
break;
}
if (signal_num != -1)
g_signal_emit (actor, actor_signals[signal_num], 0,
event, &retval);
}
out:
g_object_unref (actor);
return retval;
}
/**
* clutter_actor_set_reactive:
* @actor: a #ClutterActor
* @reactive: whether the actor should be reactive to events
*
* Sets @actor as reactive. Reactive actors will receive events.
*
* Since: 0.6
*/
void
clutter_actor_set_reactive (ClutterActor *actor,
gboolean reactive)
{
g_return_if_fail (CLUTTER_IS_ACTOR (actor));
if (reactive == CLUTTER_ACTOR_IS_REACTIVE (actor))
return;
if (reactive)
CLUTTER_ACTOR_SET_FLAGS (actor, CLUTTER_ACTOR_REACTIVE);
else
CLUTTER_ACTOR_UNSET_FLAGS (actor, CLUTTER_ACTOR_REACTIVE);
g_object_notify_by_pspec (G_OBJECT (actor), obj_props[PROP_REACTIVE]);
}
/**
* clutter_actor_get_reactive:
* @actor: a #ClutterActor
*
* Checks whether @actor is marked as reactive.
*
* Return value: %TRUE if the actor is reactive
*
* Since: 0.6
*/
gboolean
clutter_actor_get_reactive (ClutterActor *actor)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (actor), FALSE);
return CLUTTER_ACTOR_IS_REACTIVE (actor) ? TRUE : FALSE;
}
/**
* clutter_actor_get_anchor_point:
* @self: a #ClutterActor
* @anchor_x: (out): return location for the X coordinate of the anchor point
* @anchor_y: (out): return location for the Y coordinate of the anchor point
*
* Gets the current anchor point of the @actor in pixels.
*
* Since: 0.6
*/
void
clutter_actor_get_anchor_point (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat *anchor_x,
gfloat *anchor_y)
{
const ClutterTransformInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
info = _clutter_actor_get_transform_info_or_defaults (self);
clutter_anchor_coord_get_units (self, &info->anchor,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
anchor_x,
anchor_y,
NULL);
}
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_set_anchor_point:
* @self: a #ClutterActor
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* @anchor_x: X coordinate of the anchor point
* @anchor_y: Y coordinate of the anchor point
*
* Sets an anchor point for @self. The anchor point is a point in the
* coordinate space of an actor to which the actor position within its
* parent is relative; the default is (0, 0), i.e. the top-left corner
* of the actor.
*
* Since: 0.6
*/
void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_set_anchor_point (ClutterActor *self,
gfloat anchor_x,
gfloat anchor_y)
{
ClutterTransformInfo *info;
ClutterActorPrivate *priv;
gboolean changed = FALSE;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat old_anchor_x, old_anchor_y;
GObject *obj;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
obj = G_OBJECT (self);
priv = self->priv;
info = _clutter_actor_get_transform_info (self);
g_object_freeze_notify (obj);
clutter_anchor_coord_get_units (self, &info->anchor,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
&old_anchor_x,
&old_anchor_y,
NULL);
if (info->anchor.is_fractional)
g_object_notify_by_pspec (obj, obj_props[PROP_ANCHOR_GRAVITY]);
if (old_anchor_x != anchor_x)
{
g_object_notify_by_pspec (obj, obj_props[PROP_ANCHOR_X]);
changed = TRUE;
}
if (old_anchor_y != anchor_y)
{
g_object_notify_by_pspec (obj, obj_props[PROP_ANCHOR_Y]);
changed = TRUE;
}
clutter_anchor_coord_set_units (&info->anchor, anchor_x, anchor_y, 0);
if (changed)
{
priv->transform_valid = FALSE;
clutter_actor_queue_redraw (self);
}
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
g_object_thaw_notify (obj);
}
/**
* clutter_actor_get_anchor_point_gravity:
* @self: a #ClutterActor
*
* Retrieves the anchor position expressed as a #ClutterGravity. If
* the anchor point was specified using pixels or units this will
* return %CLUTTER_GRAVITY_NONE.
*
* Return value: the #ClutterGravity used by the anchor point
*
* Since: 1.0
*/
ClutterGravity
clutter_actor_get_anchor_point_gravity (ClutterActor *self)
{
const ClutterTransformInfo *info;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), CLUTTER_GRAVITY_NONE);
info = _clutter_actor_get_transform_info_or_defaults (self);
return clutter_anchor_coord_get_gravity (&info->anchor);
}
/**
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* clutter_actor_move_anchor_point:
* @self: a #ClutterActor
* @anchor_x: X coordinate of the anchor point
* @anchor_y: Y coordinate of the anchor point
*
* Sets an anchor point for the actor, and adjusts the actor postion so that
* the relative position of the actor toward its parent remains the same.
*
* Since: 0.6
*/
void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_move_anchor_point (ClutterActor *self,
gfloat anchor_x,
gfloat anchor_y)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat old_anchor_x, old_anchor_y;
const ClutterTransformInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
info = _clutter_actor_get_transform_info (self);
clutter_anchor_coord_get_units (self, &info->anchor,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
&old_anchor_x,
&old_anchor_y,
NULL);
g_object_freeze_notify (G_OBJECT (self));
clutter_actor_set_anchor_point (self, anchor_x, anchor_y);
if (self->priv->position_set)
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_move_by (self,
anchor_x - old_anchor_x,
anchor_y - old_anchor_y);
g_object_thaw_notify (G_OBJECT (self));
}
/**
* clutter_actor_move_anchor_point_from_gravity:
* @self: a #ClutterActor
* @gravity: #ClutterGravity.
*
* Sets an anchor point on the actor based on the given gravity, adjusting the
* actor postion so that its relative position within its parent remains
* unchanged.
*
* Since version 1.0 the anchor point will be stored as a gravity so
* that if the actor changes size then the anchor point will move. For
* example, if you set the anchor point to %CLUTTER_GRAVITY_SOUTH_EAST
* and later double the size of the actor, the anchor point will move
* to the bottom right.
*
* Since: 0.6
*/
void
clutter_actor_move_anchor_point_from_gravity (ClutterActor *self,
ClutterGravity gravity)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat old_anchor_x, old_anchor_y, new_anchor_x, new_anchor_y;
const ClutterTransformInfo *info;
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
info = _clutter_actor_get_transform_info (self);
g_object_freeze_notify (G_OBJECT (self));
clutter_anchor_coord_get_units (self, &info->anchor,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
&old_anchor_x,
&old_anchor_y,
NULL);
clutter_actor_set_anchor_point_from_gravity (self, gravity);
clutter_anchor_coord_get_units (self, &info->anchor,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
&new_anchor_x,
&new_anchor_y,
NULL);
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
if (priv->position_set)
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_move_by (self,
new_anchor_x - old_anchor_x,
new_anchor_y - old_anchor_y);
g_object_thaw_notify (G_OBJECT (self));
}
/**
* clutter_actor_set_anchor_point_from_gravity:
* @self: a #ClutterActor
* @gravity: #ClutterGravity.
*
* Sets an anchor point on the actor, based on the given gravity (this is a
* convenience function wrapping clutter_actor_set_anchor_point()).
*
* Since version 1.0 the anchor point will be stored as a gravity so
* that if the actor changes size then the anchor point will move. For
* example, if you set the anchor point to %CLUTTER_GRAVITY_SOUTH_EAST
* and later double the size of the actor, the anchor point will move
* to the bottom right.
*
* Since: 0.6
*/
void
clutter_actor_set_anchor_point_from_gravity (ClutterActor *self,
ClutterGravity gravity)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (gravity == CLUTTER_GRAVITY_NONE)
clutter_actor_set_anchor_point (self, 0, 0);
else
{
GObject *obj = G_OBJECT (self);
ClutterTransformInfo *info;
g_object_freeze_notify (obj);
info = _clutter_actor_get_transform_info (self);
clutter_anchor_coord_set_gravity (&info->anchor, gravity);
g_object_notify_by_pspec (obj, obj_props[PROP_ANCHOR_GRAVITY]);
g_object_notify_by_pspec (obj, obj_props[PROP_ANCHOR_X]);
g_object_notify_by_pspec (obj, obj_props[PROP_ANCHOR_Y]);
self->priv->transform_valid = FALSE;
clutter_actor_queue_redraw (self);
g_object_thaw_notify (obj);
}
}
static void
clutter_container_iface_init (ClutterContainerIface *iface)
{
/* we don't override anything, as ClutterContainer already has a default
* implementation that we can use, and which calls into our own API.
*/
}
typedef enum
{
PARSE_X,
PARSE_Y,
PARSE_WIDTH,
PARSE_HEIGHT,
PARSE_ANCHOR_X,
PARSE_ANCHOR_Y
} ParseDimension;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
static gfloat
parse_units (ClutterActor *self,
ParseDimension dimension,
JsonNode *node)
{
GValue value = { 0, };
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat retval = 0;
if (JSON_NODE_TYPE (node) != JSON_NODE_VALUE)
return 0;
json_node_get_value (node, &value);
if (G_VALUE_HOLDS (&value, G_TYPE_INT64))
{
retval = (gfloat) g_value_get_int64 (&value);
}
else if (G_VALUE_HOLDS (&value, G_TYPE_DOUBLE))
{
retval = g_value_get_double (&value);
}
else if (G_VALUE_HOLDS (&value, G_TYPE_STRING))
{
ClutterUnits units;
gboolean res;
res = clutter_units_from_string (&units, g_value_get_string (&value));
if (res)
retval = clutter_units_to_pixels (&units);
else
{
g_warning ("Invalid value '%s': integers, strings or floating point "
"values can be used for the x, y, width and height "
"properties. Valid modifiers for strings are 'px', 'mm', "
"'pt' and 'em'.",
g_value_get_string (&value));
retval = 0;
}
}
else
{
g_warning ("Invalid value of type '%s': integers, strings of floating "
"point values can be used for the x, y, width, height "
"anchor-x and anchor-y properties.",
g_type_name (G_VALUE_TYPE (&value)));
}
g_value_unset (&value);
return retval;
}
typedef struct {
ClutterRotateAxis axis;
gdouble angle;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat center_x;
gfloat center_y;
gfloat center_z;
} RotationInfo;
static inline gboolean
parse_rotation_array (ClutterActor *actor,
JsonArray *array,
RotationInfo *info)
{
JsonNode *element;
if (json_array_get_length (array) != 2)
return FALSE;
/* angle */
element = json_array_get_element (array, 0);
if (JSON_NODE_TYPE (element) == JSON_NODE_VALUE)
info->angle = json_node_get_double (element);
else
return FALSE;
/* center */
element = json_array_get_element (array, 1);
if (JSON_NODE_TYPE (element) == JSON_NODE_ARRAY)
{
JsonArray *center = json_node_get_array (element);
if (json_array_get_length (center) != 2)
return FALSE;
switch (info->axis)
{
case CLUTTER_X_AXIS:
info->center_y = parse_units (actor, PARSE_Y,
json_array_get_element (center, 0));
info->center_z = parse_units (actor, PARSE_Y,
json_array_get_element (center, 1));
return TRUE;
case CLUTTER_Y_AXIS:
info->center_x = parse_units (actor, PARSE_X,
json_array_get_element (center, 0));
info->center_z = parse_units (actor, PARSE_X,
json_array_get_element (center, 1));
return TRUE;
case CLUTTER_Z_AXIS:
info->center_x = parse_units (actor, PARSE_X,
json_array_get_element (center, 0));
info->center_y = parse_units (actor, PARSE_Y,
json_array_get_element (center, 1));
return TRUE;
}
}
return FALSE;
}
static gboolean
parse_rotation (ClutterActor *actor,
JsonNode *node,
RotationInfo *info)
{
JsonArray *array;
guint len, i;
gboolean retval = FALSE;
if (JSON_NODE_TYPE (node) != JSON_NODE_ARRAY)
{
g_warning ("Invalid node of type '%s' found, expecting an array",
json_node_type_name (node));
return FALSE;
}
array = json_node_get_array (node);
len = json_array_get_length (array);
for (i = 0; i < len; i++)
{
JsonNode *element = json_array_get_element (array, i);
JsonObject *object;
JsonNode *member;
if (JSON_NODE_TYPE (element) != JSON_NODE_OBJECT)
{
g_warning ("Invalid node of type '%s' found, expecting an object",
json_node_type_name (element));
return FALSE;
}
object = json_node_get_object (element);
if (json_object_has_member (object, "x-axis"))
{
member = json_object_get_member (object, "x-axis");
info->axis = CLUTTER_X_AXIS;
if (JSON_NODE_TYPE (member) == JSON_NODE_VALUE)
{
info->angle = json_node_get_double (member);
retval = TRUE;
}
else if (JSON_NODE_TYPE (member) == JSON_NODE_ARRAY)
retval = parse_rotation_array (actor,
json_node_get_array (member),
info);
else
retval = FALSE;
}
else if (json_object_has_member (object, "y-axis"))
{
member = json_object_get_member (object, "y-axis");
info->axis = CLUTTER_Y_AXIS;
if (JSON_NODE_TYPE (member) == JSON_NODE_VALUE)
{
info->angle = json_node_get_double (member);
retval = TRUE;
}
else if (JSON_NODE_TYPE (member) == JSON_NODE_ARRAY)
retval = parse_rotation_array (actor,
json_node_get_array (member),
info);
else
retval = FALSE;
}
else if (json_object_has_member (object, "z-axis"))
{
member = json_object_get_member (object, "z-axis");
info->axis = CLUTTER_Z_AXIS;
if (JSON_NODE_TYPE (member) == JSON_NODE_VALUE)
{
info->angle = json_node_get_double (member);
retval = TRUE;
}
else if (JSON_NODE_TYPE (member) == JSON_NODE_ARRAY)
retval = parse_rotation_array (actor,
json_node_get_array (member),
info);
else
retval = FALSE;
}
}
return retval;
}
static GSList *
parse_actor_metas (ClutterScript *script,
ClutterActor *actor,
JsonNode *node)
{
GList *elements, *l;
GSList *retval = NULL;
if (!JSON_NODE_HOLDS_ARRAY (node))
return NULL;
elements = json_array_get_elements (json_node_get_array (node));
for (l = elements; l != NULL; l = l->next)
{
JsonNode *element = l->data;
const gchar *id_ = _clutter_script_get_id_from_node (element);
GObject *meta;
if (id_ == NULL || *id_ == '\0')
continue;
meta = clutter_script_get_object (script, id_);
if (meta == NULL)
continue;
retval = g_slist_prepend (retval, meta);
}
g_list_free (elements);
return g_slist_reverse (retval);
}
static GSList *
parse_behaviours (ClutterScript *script,
ClutterActor *actor,
JsonNode *node)
{
GList *elements, *l;
GSList *retval = NULL;
if (!JSON_NODE_HOLDS_ARRAY (node))
return NULL;
elements = json_array_get_elements (json_node_get_array (node));
for (l = elements; l != NULL; l = l->next)
{
JsonNode *element = l->data;
const gchar *id_ = _clutter_script_get_id_from_node (element);
GObject *behaviour;
if (id_ == NULL || *id_ == '\0')
continue;
behaviour = clutter_script_get_object (script, id_);
if (behaviour == NULL)
continue;
retval = g_slist_prepend (retval, behaviour);
}
g_list_free (elements);
return g_slist_reverse (retval);
}
static gboolean
clutter_actor_parse_custom_node (ClutterScriptable *scriptable,
ClutterScript *script,
GValue *value,
const gchar *name,
JsonNode *node)
{
ClutterActor *actor = CLUTTER_ACTOR (scriptable);
gboolean retval = FALSE;
if ((name[0] == 'x' && name[1] == '\0') ||
(name[0] == 'y' && name[1] == '\0') ||
(strcmp (name, "width") == 0) ||
(strcmp (name, "height") == 0) ||
(strcmp (name, "anchor_x") == 0) ||
(strcmp (name, "anchor_y") == 0))
{
ParseDimension dimension;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat units;
if (name[0] == 'x')
dimension = PARSE_X;
else if (name[0] == 'y')
dimension = PARSE_Y;
else if (name[0] == 'w')
dimension = PARSE_WIDTH;
else if (name[0] == 'h')
dimension = PARSE_HEIGHT;
else if (name[0] == 'a' && name[7] == 'x')
dimension = PARSE_ANCHOR_X;
else if (name[0] == 'a' && name[7] == 'y')
dimension = PARSE_ANCHOR_Y;
else
return FALSE;
units = parse_units (actor, dimension, node);
/* convert back to pixels: all properties are pixel-based */
g_value_init (value, G_TYPE_FLOAT);
g_value_set_float (value, units);
retval = TRUE;
}
else if (strcmp (name, "rotation") == 0)
{
RotationInfo *info;
info = g_slice_new0 (RotationInfo);
retval = parse_rotation (actor, node, info);
if (retval)
{
g_value_init (value, G_TYPE_POINTER);
g_value_set_pointer (value, info);
}
else
g_slice_free (RotationInfo, info);
}
else if (strcmp (name, "behaviours") == 0)
{
GSList *l;
#ifdef CLUTTER_ENABLE_DEBUG
if (G_UNLIKELY (_clutter_diagnostic_enabled ()))
_clutter_diagnostic_message ("The 'behaviours' key is deprecated "
"and it should not be used in newly "
"written ClutterScript definitions.");
#endif
l = parse_behaviours (script, actor, node);
g_value_init (value, G_TYPE_POINTER);
g_value_set_pointer (value, l);
retval = TRUE;
}
else if (strcmp (name, "actions") == 0 ||
strcmp (name, "constraints") == 0 ||
strcmp (name, "effects") == 0)
{
GSList *l;
l = parse_actor_metas (script, actor, node);
g_value_init (value, G_TYPE_POINTER);
g_value_set_pointer (value, l);
retval = TRUE;
}
return retval;
}
static void
clutter_actor_set_custom_property (ClutterScriptable *scriptable,
ClutterScript *script,
const gchar *name,
const GValue *value)
{
ClutterActor *actor = CLUTTER_ACTOR (scriptable);
#ifdef CLUTTER_ENABLE_DEBUG
if (G_UNLIKELY (CLUTTER_HAS_DEBUG (SCRIPT)))
{
gchar *tmp = g_strdup_value_contents (value);
CLUTTER_NOTE (SCRIPT,
"in ClutterActor::set_custom_property('%s') = %s",
name,
tmp);
g_free (tmp);
}
#endif /* CLUTTER_ENABLE_DEBUG */
if (strcmp (name, "rotation") == 0)
{
RotationInfo *info;
if (!G_VALUE_HOLDS (value, G_TYPE_POINTER))
return;
info = g_value_get_pointer (value);
clutter_actor_set_rotation (actor,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
info->axis, info->angle,
info->center_x,
info->center_y,
info->center_z);
g_slice_free (RotationInfo, info);
return;
}
if (strcmp (name, "behaviours") == 0)
{
GSList *behaviours, *l;
if (!G_VALUE_HOLDS (value, G_TYPE_POINTER))
return;
behaviours = g_value_get_pointer (value);
for (l = behaviours; l != NULL; l = l->next)
{
ClutterBehaviour *behaviour = l->data;
clutter_behaviour_apply (behaviour, actor);
}
g_slist_free (behaviours);
return;
}
if (strcmp (name, "actions") == 0 ||
strcmp (name, "constraints") == 0 ||
strcmp (name, "effects") == 0)
{
GSList *metas, *l;
if (!G_VALUE_HOLDS (value, G_TYPE_POINTER))
return;
metas = g_value_get_pointer (value);
for (l = metas; l != NULL; l = l->next)
{
if (name[0] == 'a')
clutter_actor_add_action (actor, l->data);
if (name[0] == 'c')
clutter_actor_add_constraint (actor, l->data);
if (name[0] == 'e')
clutter_actor_add_effect (actor, l->data);
}
g_slist_free (metas);
return;
}
g_object_set_property (G_OBJECT (scriptable), name, value);
}
static void
clutter_scriptable_iface_init (ClutterScriptableIface *iface)
{
iface->parse_custom_node = clutter_actor_parse_custom_node;
iface->set_custom_property = clutter_actor_set_custom_property;
}
static ClutterActorMeta *
get_meta_from_animation_property (ClutterActor *actor,
const gchar *name,
gchar **name_p)
{
ClutterActorPrivate *priv = actor->priv;
ClutterActorMeta *meta = NULL;
gchar **tokens;
/* if this is not a special property, fall through */
if (name[0] != '@')
return NULL;
/* detect the properties named using the following spec:
*
* @<section>.<meta-name>.<property-name>
*
* where <section> can be one of the following:
*
* - actions
* - constraints
* - effects
*
* and <meta-name> is the name set on a specific ActorMeta
*/
tokens = g_strsplit (name + 1, ".", -1);
if (tokens == NULL || g_strv_length (tokens) != 3)
{
CLUTTER_NOTE (ANIMATION, "Invalid property name '%s'",
name + 1);
g_strfreev (tokens);
return NULL;
}
if (strcmp (tokens[0], "actions") == 0)
meta = _clutter_meta_group_get_meta (priv->actions, tokens[1]);
if (strcmp (tokens[0], "constraints") == 0)
meta = _clutter_meta_group_get_meta (priv->constraints, tokens[1]);
if (strcmp (tokens[0], "effects") == 0)
meta = _clutter_meta_group_get_meta (priv->effects, tokens[1]);
if (name_p != NULL)
*name_p = g_strdup (tokens[2]);
CLUTTER_NOTE (ANIMATION,
"Looking for property '%s' of object '%s' in section '%s'",
tokens[2],
tokens[1],
tokens[0]);
g_strfreev (tokens);
return meta;
}
static GParamSpec *
clutter_actor_find_property (ClutterAnimatable *animatable,
const gchar *property_name)
{
ClutterActorMeta *meta = NULL;
GObjectClass *klass = NULL;
GParamSpec *pspec = NULL;
gchar *p_name = NULL;
meta = get_meta_from_animation_property (CLUTTER_ACTOR (animatable),
property_name,
&p_name);
if (meta != NULL)
{
klass = G_OBJECT_GET_CLASS (meta);
pspec = g_object_class_find_property (klass, p_name);
}
else
{
klass = G_OBJECT_GET_CLASS (animatable);
pspec = g_object_class_find_property (klass, property_name);
}
g_free (p_name);
return pspec;
}
static void
clutter_actor_get_initial_state (ClutterAnimatable *animatable,
const gchar *property_name,
GValue *initial)
{
ClutterActorMeta *meta = NULL;
gchar *p_name = NULL;
meta = get_meta_from_animation_property (CLUTTER_ACTOR (animatable),
property_name,
&p_name);
if (meta != NULL)
g_object_get_property (G_OBJECT (meta), p_name, initial);
else
g_object_get_property (G_OBJECT (animatable), property_name, initial);
g_free (p_name);
}
static void
clutter_actor_set_final_state (ClutterAnimatable *animatable,
const gchar *property_name,
const GValue *final)
{
ClutterActorMeta *meta = NULL;
gchar *p_name = NULL;
meta = get_meta_from_animation_property (CLUTTER_ACTOR (animatable),
property_name,
&p_name);
if (meta != NULL)
g_object_set_property (G_OBJECT (meta), p_name, final);
else
g_object_set_property (G_OBJECT (animatable), property_name, final);
g_free (p_name);
}
static void
clutter_animatable_iface_init (ClutterAnimatableIface *iface)
{
iface->find_property = clutter_actor_find_property;
iface->get_initial_state = clutter_actor_get_initial_state;
iface->set_final_state = clutter_actor_set_final_state;
}
/**
* clutter_actor_transform_stage_point:
* @self: A #ClutterActor
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
* @x: (in): x screen coordinate of the point to unproject
* @y: (in): y screen coordinate of the point to unproject
* @x_out: (out): return location for the unprojected x coordinance
* @y_out: (out): return location for the unprojected y coordinance
*
* This function translates screen coordinates (@x, @y) to
* coordinates relative to the actor. For example, it can be used to translate
* screen events from global screen coordinates into actor-local coordinates.
*
* The conversion can fail, notably if the transform stack results in the
* actor being projected on the screen as a mere line.
*
* The conversion should not be expected to be pixel-perfect due to the
* nature of the operation. In general the error grows when the skewing
* of the actor rectangle on screen increases.
*
* <note><para>This function can be computationally intensive.</para></note>
*
* <note><para>This function only works when the allocation is up-to-date,
* i.e. inside of paint().</para></note>
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
*
* Return value: %TRUE if conversion was successful.
*
* Since: 0.6
*/
gboolean
clutter_actor_transform_stage_point (ClutterActor *self,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat x,
gfloat y,
gfloat *x_out,
gfloat *y_out)
{
ClutterVertex v[4];
float ST[3][3];
float RQ[3][3];
int du, dv, xi, yi;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
float px, py;
float xf, yf, wf, det;
ClutterActorPrivate *priv;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
priv = self->priv;
/* This implementation is based on the quad -> quad projection algorithm
* described by Paul Heckbert in:
*
* http://www.cs.cmu.edu/~ph/texfund/texfund.pdf
*
* and the sample implementation at:
*
* http://www.cs.cmu.edu/~ph/src/texfund/
*
* Our texture is a rectangle with origin [0, 0], so we are mapping from
* quad to rectangle only, which significantly simplifies things; the
* function calls have been unrolled, and most of the math is done in fixed
* point.
*/
2008-06-10 Emmanuele Bassi <ebassi@openedhand.com> Bug #815 - Split up request, allocation, and paint box * clutter/clutter-actor.[ch]: Rework the size allocation, request and paint area. Now ::request_coords() is called ::allocate(), and ::query_coords() has been split into ::get_preferred_width() and ::get_preferred_height(). See the documentation and the layout test on how to implement a container and layout manager with the new API. (#915, based on a patch by Havoc Pennington, Lucas Rocha and Johan Bilien) * clutter/clutter-clone-texture.c: Port CloneTexture to the new size negotiation API; it just means forwarding the requests to the parent texture. * clutter/clutter-deprecated.h: Add deprecated and replaced API. * clutter/clutter-entry.c: Port Entry to the new size negotiation API. * clutter/clutter-group.c: Port Group to the new size negotiation API; the semantics of the Group actor do not change. * clutter/clutter-label.c: Port Label to the new size negotiation API, and vastly simplify the code. * clutter/clutter-main.[ch]: Add API for executing a relayout when needed. * clutter/clutter-private.h: Add new Stage private API. * clutter/clutter-rectangle.c: Update the get_abs_opacity() call to get_paint_opacity(). * clutter/clutter-stage.c: (clutter_stage_get_preferred_width), (clutter_stage_get_preferred_height), (clutter_stage_allocate), (clutter_stage_class_init): Port Stage to the new size negotiation API. * clutter/clutter-texture.c: Port Texture to the new size negotiation API. * clutter/clutter-types.h: Add ClutterRequestMode enumeration. * clutter/x11/clutter-stage-x11.c: Port the X11 stage implementation to the new size negotiation API. * tests/Makefile.am: Add the layout manager test case. * tests/test-opacity.c: Update. * tests/test-project.c: Update. * tests/test-layout.c: Test case for a layout manager implemented using the new size negotiation API; the layout manager handles both transformed and untransformed children.
2008-06-10 17:07:52 +00:00
clutter_actor_get_abs_allocation_vertices (self, v);
/* Keeping these as ints simplifies the multiplication (no significant
* loss of precision here).
*/
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
du = (int) (priv->allocation.x2 - priv->allocation.x1);
dv = (int) (priv->allocation.y2 - priv->allocation.y1);
if (!du || !dv)
return FALSE;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
#define UX2FP(x) (x)
#define DET2FP(a,b,c,d) (((a) * (d)) - ((b) * (c)))
/* First, find mapping from unit uv square to xy quadrilateral; this
* equivalent to the pmap_square_quad() functions in the sample
* implementation, which we can simplify, since our target is always
* a rectangle.
*/
px = v[0].x - v[1].x + v[3].x - v[2].x;
py = v[0].y - v[1].y + v[3].y - v[2].y;
if (!px && !py)
{
/* affine transform */
RQ[0][0] = UX2FP (v[1].x - v[0].x);
RQ[1][0] = UX2FP (v[3].x - v[1].x);
RQ[2][0] = UX2FP (v[0].x);
RQ[0][1] = UX2FP (v[1].y - v[0].y);
RQ[1][1] = UX2FP (v[3].y - v[1].y);
RQ[2][1] = UX2FP (v[0].y);
RQ[0][2] = 0;
RQ[1][2] = 0;
RQ[2][2] = 1.0;
}
else
{
/* projective transform */
double dx1, dx2, dy1, dy2, del;
dx1 = UX2FP (v[1].x - v[3].x);
dx2 = UX2FP (v[2].x - v[3].x);
dy1 = UX2FP (v[1].y - v[3].y);
dy2 = UX2FP (v[2].y - v[3].y);
del = DET2FP (dx1, dx2, dy1, dy2);
if (!del)
return FALSE;
/*
* The division here needs to be done in floating point for
* precisions reasons.
*/
RQ[0][2] = (DET2FP (UX2FP (px), dx2, UX2FP (py), dy2) / del);
RQ[1][2] = (DET2FP (dx1, UX2FP (px), dy1, UX2FP (py)) / del);
RQ[1][2] = (DET2FP (dx1, UX2FP (px), dy1, UX2FP (py)) / del);
RQ[2][2] = 1.0;
RQ[0][0] = UX2FP (v[1].x - v[0].x) + (RQ[0][2] * UX2FP (v[1].x));
RQ[1][0] = UX2FP (v[2].x - v[0].x) + (RQ[1][2] * UX2FP (v[2].x));
RQ[2][0] = UX2FP (v[0].x);
RQ[0][1] = UX2FP (v[1].y - v[0].y) + (RQ[0][2] * UX2FP (v[1].y));
RQ[1][1] = UX2FP (v[2].y - v[0].y) + (RQ[1][2] * UX2FP (v[2].y));
RQ[2][1] = UX2FP (v[0].y);
}
/*
* Now combine with transform from our rectangle (u0,v0,u1,v1) to unit
* square. Since our rectangle is based at 0,0 we only need to scale.
*/
RQ[0][0] /= du;
RQ[1][0] /= dv;
RQ[0][1] /= du;
RQ[1][1] /= dv;
RQ[0][2] /= du;
RQ[1][2] /= dv;
/*
* Now RQ is transform from uv rectangle to xy quadrilateral; we need an
* inverse of that.
*/
ST[0][0] = DET2FP (RQ[1][1], RQ[1][2], RQ[2][1], RQ[2][2]);
ST[1][0] = DET2FP (RQ[1][2], RQ[1][0], RQ[2][2], RQ[2][0]);
ST[2][0] = DET2FP (RQ[1][0], RQ[1][1], RQ[2][0], RQ[2][1]);
ST[0][1] = DET2FP (RQ[2][1], RQ[2][2], RQ[0][1], RQ[0][2]);
ST[1][1] = DET2FP (RQ[2][2], RQ[2][0], RQ[0][2], RQ[0][0]);
ST[2][1] = DET2FP (RQ[2][0], RQ[2][1], RQ[0][0], RQ[0][1]);
ST[0][2] = DET2FP (RQ[0][1], RQ[0][2], RQ[1][1], RQ[1][2]);
ST[1][2] = DET2FP (RQ[0][2], RQ[0][0], RQ[1][2], RQ[1][0]);
ST[2][2] = DET2FP (RQ[0][0], RQ[0][1], RQ[1][0], RQ[1][1]);
/*
* Check the resulting matrix is OK.
*/
det = (RQ[0][0] * ST[0][0])
+ (RQ[0][1] * ST[0][1])
+ (RQ[0][2] * ST[0][2]);
if (!det)
return FALSE;
/*
* Now transform our point with the ST matrix; the notional w
* coordinate is 1, hence the last part is simply added.
*/
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
xi = (int) x;
yi = (int) y;
xf = xi * ST[0][0] + yi * ST[1][0] + ST[2][0];
yf = xi * ST[0][1] + yi * ST[1][1] + ST[2][1];
wf = xi * ST[0][2] + yi * ST[1][2] + ST[2][2];
if (x_out)
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
*x_out = xf / wf;
if (y_out)
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
*y_out = yf / wf;
#undef UX2FP
#undef DET2FP
return TRUE;
}
/*
* ClutterGeometry
*/
static ClutterGeometry*
clutter_geometry_copy (const ClutterGeometry *geometry)
{
return g_slice_dup (ClutterGeometry, geometry);
}
static void
clutter_geometry_free (ClutterGeometry *geometry)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (G_LIKELY (geometry != NULL))
g_slice_free (ClutterGeometry, geometry);
}
/**
* clutter_geometry_union:
* @geometry_a: a #ClutterGeometry
* @geometry_b: another #ClutterGeometry
* @result: (out): location to store the result
*
* Find the union of two rectangles represented as #ClutterGeometry.
*
* Since: 1.4
*/
void
clutter_geometry_union (const ClutterGeometry *geometry_a,
const ClutterGeometry *geometry_b,
ClutterGeometry *result)
{
/* We don't try to handle rectangles that can't be represented
* as a signed integer box */
gint x_1 = MIN (geometry_a->x, geometry_b->x);
gint y_1 = MIN (geometry_a->y, geometry_b->y);
gint x_2 = MAX (geometry_a->x + (gint)geometry_a->width,
geometry_b->x + (gint)geometry_b->width);
gint y_2 = MAX (geometry_a->y + (gint)geometry_a->height,
geometry_b->y + (gint)geometry_b->height);
result->x = x_1;
result->y = y_1;
result->width = x_2 - x_1;
result->height = y_2 - y_1;
}
/**
* clutter_geometry_intersects:
* @geometry0: The first geometry to test
* @geometry1: The second geometry to test
*
* Determines if @geometry0 and geometry1 intersect returning %TRUE if
* they do else %FALSE.
*
* Return value: %TRUE of @geometry0 and geometry1 intersect else
* %FALSE.
*
* Since: 1.4
*/
gboolean
clutter_geometry_intersects (const ClutterGeometry *geometry0,
const ClutterGeometry *geometry1)
{
if (geometry1->x >= (geometry0->x + (gint)geometry0->width) ||
geometry1->y >= (geometry0->y + (gint)geometry0->height) ||
(geometry1->x + (gint)geometry1->width) <= geometry0->x ||
(geometry1->y + (gint)geometry1->height) <= geometry0->y)
return FALSE;
else
return TRUE;
}
static gboolean
clutter_geometry_progress (const GValue *a,
const GValue *b,
gdouble progress,
GValue *retval)
{
const ClutterGeometry *a_geom = g_value_get_boxed (a);
const ClutterGeometry *b_geom = g_value_get_boxed (b);
ClutterGeometry res = { 0, };
gint a_width = a_geom->width;
gint b_width = b_geom->width;
gint a_height = a_geom->height;
gint b_height = b_geom->height;
res.x = a_geom->x + (b_geom->x - a_geom->x) * progress;
res.y = a_geom->y + (b_geom->y - a_geom->y) * progress;
res.width = a_width + (b_width - a_width) * progress;
res.height = a_height + (b_height - a_height) * progress;
g_value_set_boxed (retval, &res);
return TRUE;
}
G_DEFINE_BOXED_TYPE_WITH_CODE (ClutterGeometry, clutter_geometry,
clutter_geometry_copy,
clutter_geometry_free,
CLUTTER_REGISTER_INTERVAL_PROGRESS (clutter_geometry_progress));
/*
* ClutterVertices
*/
/**
* clutter_vertex_new:
* @x: X coordinate
* @y: Y coordinate
* @z: Z coordinate
*
* Creates a new #ClutterVertex for the point in 3D space
* identified by the 3 coordinates @x, @y, @z
*
* Return value: the newly allocate #ClutterVertex. Use
* clutter_vertex_free() to free the resources
*
* Since: 1.0
*/
ClutterVertex *
clutter_vertex_new (gfloat x,
gfloat y,
gfloat z)
{
ClutterVertex *vertex;
vertex = g_slice_new (ClutterVertex);
vertex->x = x;
vertex->y = y;
vertex->z = z;
return vertex;
}
/**
* clutter_vertex_copy:
* @vertex: a #ClutterVertex
*
* Copies @vertex
*
* Return value: a newly allocated copy of #ClutterVertex. Use
* clutter_vertex_free() to free the allocated resources
*
* Since: 1.0
*/
ClutterVertex *
2007-06-12 11:42:29 +00:00
clutter_vertex_copy (const ClutterVertex *vertex)
{
if (G_LIKELY (vertex != NULL))
return g_slice_dup (ClutterVertex, vertex);
return NULL;
}
/**
* clutter_vertex_free:
* @vertex: a #ClutterVertex
*
* Frees a #ClutterVertex allocated using clutter_vertex_copy()
*
* Since: 1.0
*/
void
clutter_vertex_free (ClutterVertex *vertex)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (G_UNLIKELY (vertex != NULL))
g_slice_free (ClutterVertex, vertex);
}
/**
* clutter_vertex_equal:
* @vertex_a: a #ClutterVertex
* @vertex_b: a #ClutterVertex
*
* Compares @vertex_a and @vertex_b for equality
*
* Return value: %TRUE if the passed #ClutterVertex are equal
*
* Since: 1.0
*/
gboolean
clutter_vertex_equal (const ClutterVertex *vertex_a,
const ClutterVertex *vertex_b)
{
g_return_val_if_fail (vertex_a != NULL && vertex_b != NULL, FALSE);
if (vertex_a == vertex_b)
return TRUE;
return vertex_a->x == vertex_b->x &&
vertex_a->y == vertex_b->y &&
vertex_a->z == vertex_b->z;
}
static gboolean
clutter_vertex_progress (const GValue *a,
const GValue *b,
gdouble progress,
GValue *retval)
{
const ClutterVertex *av = g_value_get_boxed (a);
const ClutterVertex *bv = g_value_get_boxed (b);
ClutterVertex res = { 0, };
res.x = av->x + (bv->x - av->x) * progress;
res.y = av->y + (bv->y - av->y) * progress;
res.z = av->z + (bv->z - av->z) * progress;
g_value_set_boxed (retval, &res);
return TRUE;
}
G_DEFINE_BOXED_TYPE_WITH_CODE (ClutterVertex, clutter_vertex,
clutter_vertex_copy,
clutter_vertex_free,
CLUTTER_REGISTER_INTERVAL_PROGRESS (clutter_vertex_progress));
/**
* clutter_actor_is_rotated:
* @self: a #ClutterActor
*
* Checks whether any rotation is applied to the actor.
*
* Return value: %TRUE if the actor is rotated.
*
* Since: 0.6
*/
gboolean
clutter_actor_is_rotated (ClutterActor *self)
{
const ClutterTransformInfo *info;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
info = _clutter_actor_get_transform_info_or_defaults (self);
if (info->rx_angle || info->ry_angle || info->rz_angle)
return TRUE;
return FALSE;
}
/**
* clutter_actor_is_scaled:
* @self: a #ClutterActor
*
* Checks whether the actor is scaled in either dimension.
*
* Return value: %TRUE if the actor is scaled.
*
* Since: 0.6
*/
gboolean
clutter_actor_is_scaled (ClutterActor *self)
{
const ClutterTransformInfo *info;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
info = _clutter_actor_get_transform_info_or_defaults (self);
if (info->scale_x != 1.0 || info->scale_y != 1.0)
return TRUE;
return FALSE;
}
ClutterActor *
_clutter_actor_get_stage_internal (ClutterActor *actor)
{
while (actor && !CLUTTER_ACTOR_IS_TOPLEVEL (actor))
actor = actor->priv->parent;
return actor;
}
/**
* clutter_actor_get_stage:
* @actor: a #ClutterActor
*
* Retrieves the #ClutterStage where @actor is contained.
*
* Return value: (transfer none) (type Clutter.Stage): the stage
* containing the actor, or %NULL
*
* Since: 0.8
*/
ClutterActor *
clutter_actor_get_stage (ClutterActor *actor)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (actor), NULL);
return _clutter_actor_get_stage_internal (actor);
}
/**
* clutter_actor_allocate_available_size:
* @self: a #ClutterActor
* @x: the actor's X coordinate
* @y: the actor's Y coordinate
* @available_width: the maximum available width, or -1 to use the
* actor's natural width
* @available_height: the maximum available height, or -1 to use the
* actor's natural height
* @flags: flags controlling the allocation
*
* Allocates @self taking into account the #ClutterActor<!-- -->'s
* preferred size, but limiting it to the maximum available width
* and height provided.
*
* This function will do the right thing when dealing with the
* actor's request mode.
*
* The implementation of this function is equivalent to:
*
* |[
* if (request_mode == CLUTTER_REQUEST_HEIGHT_FOR_WIDTH)
* {
* clutter_actor_get_preferred_width (self, available_height,
* &amp;min_width,
* &amp;natural_width);
* width = CLAMP (natural_width, min_width, available_width);
*
* clutter_actor_get_preferred_height (self, width,
* &amp;min_height,
* &amp;natural_height);
* height = CLAMP (natural_height, min_height, available_height);
* }
* else
* {
* clutter_actor_get_preferred_height (self, available_width,
* &amp;min_height,
* &amp;natural_height);
* height = CLAMP (natural_height, min_height, available_height);
*
* clutter_actor_get_preferred_width (self, height,
* &amp;min_width,
* &amp;natural_width);
* width = CLAMP (natural_width, min_width, available_width);
* }
*
* box.x1 = x; box.y1 = y;
* box.x2 = box.x1 + available_width;
* box.y2 = box.y1 + available_height;
* clutter_actor_allocate (self, &amp;box, flags);
* ]|
*
* This function can be used by fluid layout managers to allocate
* an actor's preferred size without making it bigger than the area
* available for the container.
*
* Since: 1.0
*/
void
clutter_actor_allocate_available_size (ClutterActor *self,
gfloat x,
gfloat y,
gfloat available_width,
gfloat available_height,
ClutterAllocationFlags flags)
{
ClutterActorPrivate *priv;
gfloat width, height;
gfloat min_width, min_height;
gfloat natural_width, natural_height;
ClutterActorBox box;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
2009-06-06 13:37:41 +00:00
width = height = 0.0;
switch (priv->request_mode)
{
case CLUTTER_REQUEST_HEIGHT_FOR_WIDTH:
clutter_actor_get_preferred_width (self, available_height,
&min_width,
&natural_width);
width = CLAMP (natural_width, min_width, available_width);
clutter_actor_get_preferred_height (self, width,
&min_height,
&natural_height);
height = CLAMP (natural_height, min_height, available_height);
break;
case CLUTTER_REQUEST_WIDTH_FOR_HEIGHT:
clutter_actor_get_preferred_height (self, available_width,
&min_height,
&natural_height);
height = CLAMP (natural_height, min_height, available_height);
clutter_actor_get_preferred_width (self, height,
&min_width,
&natural_width);
width = CLAMP (natural_width, min_width, available_width);
break;
}
box.x1 = x;
box.y1 = y;
box.x2 = box.x1 + width;
box.y2 = box.y1 + height;
clutter_actor_allocate (self, &box, flags);
}
/**
* clutter_actor_allocate_preferred_size:
* @self: a #ClutterActor
* @flags: flags controlling the allocation
*
* Allocates the natural size of @self.
*
* This function is a utility call for #ClutterActor implementations
* that allocates the actor's preferred natural size. It can be used
* by fixed layout managers (like #ClutterGroup or so called
* 'composite actors') inside the ClutterActor::allocate
* implementation to give each child exactly how much space it
* requires.
*
* This function is not meant to be used by applications. It is also
* not meant to be used outside the implementation of the
* ClutterActor::allocate virtual function.
*
* Since: 0.8
*/
void
clutter_actor_allocate_preferred_size (ClutterActor *self,
ClutterAllocationFlags flags)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat actor_x, actor_y;
gfloat natural_width, natural_height;
ClutterActorBox actor_box;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
actor_x = clutter_actor_get_x (self);
actor_y = clutter_actor_get_y (self);
clutter_actor_get_preferred_size (self,
NULL, NULL,
&natural_width,
&natural_height);
actor_box.x1 = actor_x;
actor_box.y1 = actor_y;
actor_box.x2 = actor_box.x1 + natural_width;
actor_box.y2 = actor_box.y1 + natural_height;
clutter_actor_allocate (self, &actor_box, flags);
}
/**
* clutter_actor_allocate_align_fill:
* @self: a #ClutterActor
* @box: a #ClutterActorBox, containing the available width and height
* @x_align: the horizontal alignment, between 0 and 1
* @y_align: the vertical alignment, between 0 and 1
* @x_fill: whether the actor should fill horizontally
* @y_fill: whether the actor should fill vertically
* @flags: allocation flags to be passed to clutter_actor_allocate()
*
* Allocates @self by taking into consideration the available allocation
* area; an alignment factor on either axis; and whether the actor should
* fill the allocation on either axis.
*
* The @box should contain the available allocation width and height;
* if the x1 and y1 members of #ClutterActorBox are not set to 0, the
* allocation will be offset by their value.
*
* This function takes into consideration the geometry request specified by
* the #ClutterActor:request-mode property, and the text direction.
*
* This function is useful for fluid layout managers, like #ClutterBinLayout
* or #ClutterTableLayout
*
* Since: 1.4
*/
void
clutter_actor_allocate_align_fill (ClutterActor *self,
const ClutterActorBox *box,
gdouble x_align,
gdouble y_align,
gboolean x_fill,
gboolean y_fill,
ClutterAllocationFlags flags)
{
ClutterActorPrivate *priv;
ClutterActorBox allocation = { 0, };
gfloat x_offset, y_offset;
gfloat available_width, available_height;
gfloat child_width, child_height;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (box != NULL);
g_return_if_fail (x_align >= 0.0 && x_align <= 1.0);
g_return_if_fail (y_align >= 0.0 && y_align <= 1.0);
priv = self->priv;
clutter_actor_box_get_origin (box, &x_offset, &y_offset);
clutter_actor_box_get_size (box, &available_width, &available_height);
if (available_width < 0)
available_width = 0;
if (available_height < 0)
available_height = 0;
if (x_fill)
{
allocation.x1 = x_offset;
allocation.x2 = allocation.x1 + available_width;
}
if (y_fill)
{
allocation.y1 = y_offset;
allocation.y2 = allocation.y1 + available_height;
}
/* if we are filling horizontally and vertically then we're done */
if (x_fill && y_fill)
goto out;
child_width = child_height = 0.0f;
if (priv->request_mode == CLUTTER_REQUEST_HEIGHT_FOR_WIDTH)
{
gfloat min_width, natural_width;
gfloat min_height, natural_height;
clutter_actor_get_preferred_width (self, available_height,
&min_width,
&natural_width);
child_width = CLAMP (natural_width, min_width, available_width);
if (!y_fill)
{
clutter_actor_get_preferred_height (self, child_width,
&min_height,
&natural_height);
child_height = CLAMP (natural_height, min_height, available_height);
}
}
else
{
gfloat min_width, natural_width;
gfloat min_height, natural_height;
clutter_actor_get_preferred_height (self, available_width,
&min_height,
&natural_height);
child_height = CLAMP (natural_height, min_height, available_height);
if (!x_fill)
{
clutter_actor_get_preferred_width (self, child_height,
&min_width,
&natural_width);
child_width = CLAMP (natural_width, min_width, available_width);
}
}
/* invert the horizontal alignment for RTL languages */
if (priv->text_direction == CLUTTER_TEXT_DIRECTION_RTL)
x_align = 1.0 - x_align;
if (!x_fill)
{
allocation.x1 = x_offset
+ ((available_width - child_width) * x_align);
allocation.x2 = allocation.x1 + child_width;
}
if (!y_fill)
{
allocation.y1 = y_offset
+ ((available_height - child_height) * y_align);
allocation.y2 = allocation.y1 + child_height;
}
out:
clutter_actor_box_clamp_to_pixel (&allocation);
clutter_actor_allocate (self, &allocation, flags);
}
/**
* clutter_actor_grab_key_focus:
* @self: a #ClutterActor
*
* Sets the key focus of the #ClutterStage including @self
* to this #ClutterActor.
*
* Since: 1.0
*/
void
clutter_actor_grab_key_focus (ClutterActor *self)
{
ClutterActor *stage;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
stage = _clutter_actor_get_stage_internal (self);
if (stage != NULL)
clutter_stage_set_key_focus (CLUTTER_STAGE (stage), self);
}
/**
* clutter_actor_get_pango_context:
* @self: a #ClutterActor
*
* Retrieves the #PangoContext for @self. The actor's #PangoContext
* is already configured using the appropriate font map, resolution
* and font options.
*
* Unlike clutter_actor_create_pango_context(), this context is owend
* by the #ClutterActor and it will be updated each time the options
* stored by the #ClutterBackend change.
*
* You can use the returned #PangoContext to create a #PangoLayout
* and render text using cogl_pango_render_layout() to reuse the
* glyphs cache also used by Clutter.
*
* Return value: (transfer none): the #PangoContext for a #ClutterActor.
* The returned #PangoContext is owned by the actor and should not be
* unreferenced by the application code
*
* Since: 1.0
*/
PangoContext *
clutter_actor_get_pango_context (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
priv = self->priv;
if (priv->pango_context != NULL)
return priv->pango_context;
priv->pango_context = _clutter_context_get_pango_context ();
g_object_ref (priv->pango_context);
return priv->pango_context;
}
/**
* clutter_actor_create_pango_context:
* @self: a #ClutterActor
*
* Creates a #PangoContext for the given actor. The #PangoContext
* is already configured using the appropriate font map, resolution
* and font options.
*
* See also clutter_actor_get_pango_context().
*
* Return value: (transfer full): the newly created #PangoContext.
* Use g_object_unref() on the returned value to deallocate its
* resources
*
* Since: 1.0
*/
PangoContext *
clutter_actor_create_pango_context (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
return _clutter_context_create_pango_context ();
}
/**
* clutter_actor_create_pango_layout:
* @self: a #ClutterActor
2009-07-02 13:21:16 +00:00
* @text: (allow-none) the text to set on the #PangoLayout, or %NULL
*
* Creates a new #PangoLayout from the same #PangoContext used
* by the #ClutterActor. The #PangoLayout is already configured
* with the font map, resolution and font options, and the
* given @text.
*
* If you want to keep around a #PangoLayout created by this
* function you will have to connect to the #ClutterBackend::font-changed
* and #ClutterBackend::resolution-changed signals, and call
* pango_layout_context_changed() in response to them.
*
* Return value: (transfer full): the newly created #PangoLayout.
* Use g_object_unref() when done
*
* Since: 1.0
*/
PangoLayout *
clutter_actor_create_pango_layout (ClutterActor *self,
const gchar *text)
{
PangoContext *context;
PangoLayout *layout;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
context = clutter_actor_get_pango_context (self);
layout = pango_layout_new (context);
if (text)
pango_layout_set_text (layout, text, -1);
return layout;
}
/* Allows overriding the calculated paint opacity. Used by ClutterClone and
* ClutterOffscreenEffect.
*/
void
_clutter_actor_set_opacity_override (ClutterActor *self,
gint opacity)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
self->priv->opacity_override = opacity;
}
gint
_clutter_actor_get_opacity_override (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), -1);
return self->priv->opacity_override;
}
/* Allows you to disable applying the actors model view transform during
* a paint. Used by ClutterClone. */
void
_clutter_actor_set_enable_model_view_transform (ClutterActor *self,
gboolean enable)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
self->priv->enable_model_view_transform = enable;
}
void
_clutter_actor_set_enable_paint_unmapped (ClutterActor *self,
gboolean enable)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
priv->enable_paint_unmapped = enable;
if (priv->enable_paint_unmapped)
{
/* Make sure that the parents of the widget are realized first;
* otherwise checks in clutter_actor_update_map_state() will
* fail.
*/
clutter_actor_realize (self);
clutter_actor_update_map_state (self, MAP_STATE_MAKE_MAPPED);
}
else
{
clutter_actor_update_map_state (self, MAP_STATE_MAKE_UNMAPPED);
}
}
static void
clutter_anchor_coord_get_units (ClutterActor *self,
const AnchorCoord *coord,
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat *x,
gfloat *y,
gfloat *z)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (coord->is_fractional)
{
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
gfloat actor_width, actor_height;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_actor_get_size (self, &actor_width, &actor_height);
if (x)
*x = actor_width * coord->v.fraction.x;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (y)
*y = actor_height * coord->v.fraction.y;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (z)
*z = 0;
}
else
{
if (x)
*x = coord->v.units.x;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (y)
*y = coord->v.units.y;
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
if (z)
*z = coord->v.units.z;
}
}
static void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_anchor_coord_set_units (AnchorCoord *coord,
gfloat x,
gfloat y,
gfloat z)
{
coord->is_fractional = FALSE;
coord->v.units.x = x;
coord->v.units.y = y;
coord->v.units.z = z;
}
static ClutterGravity
clutter_anchor_coord_get_gravity (const AnchorCoord *coord)
{
if (coord->is_fractional)
{
if (coord->v.fraction.x == 0.0)
{
if (coord->v.fraction.y == 0.0)
return CLUTTER_GRAVITY_NORTH_WEST;
else if (coord->v.fraction.y == 0.5)
return CLUTTER_GRAVITY_WEST;
else if (coord->v.fraction.y == 1.0)
return CLUTTER_GRAVITY_SOUTH_WEST;
else
return CLUTTER_GRAVITY_NONE;
}
else if (coord->v.fraction.x == 0.5)
{
if (coord->v.fraction.y == 0.0)
return CLUTTER_GRAVITY_NORTH;
else if (coord->v.fraction.y == 0.5)
return CLUTTER_GRAVITY_CENTER;
else if (coord->v.fraction.y == 1.0)
return CLUTTER_GRAVITY_SOUTH;
else
return CLUTTER_GRAVITY_NONE;
}
else if (coord->v.fraction.x == 1.0)
{
if (coord->v.fraction.y == 0.0)
return CLUTTER_GRAVITY_NORTH_EAST;
else if (coord->v.fraction.y == 0.5)
return CLUTTER_GRAVITY_EAST;
else if (coord->v.fraction.y == 1.0)
return CLUTTER_GRAVITY_SOUTH_EAST;
else
return CLUTTER_GRAVITY_NONE;
}
else
return CLUTTER_GRAVITY_NONE;
}
else
return CLUTTER_GRAVITY_NONE;
}
static void
Remove Units from the public API With the recent change to internal floating point values, ClutterUnit has become a redundant type, defined to be a float. All integer entry points are being internally converted to floating point values to be passed to the GL pipeline with the least amount of conversion. ClutterUnit is thus exposed as just a "pixel with fractionary bits", and not -- as users might think -- as generic, resolution and device independent units. not that it was the case, but a definitive amount of people was convinced it did provide this "feature", and was flummoxed about the mere existence of this type. So, having ClutterUnit exposed in the public API doubles the entry points and has the following disadvantages: - we have to maintain twice the amount of entry points in ClutterActor - we still do an integer-to-float implicit conversion - we introduce a weird impedance between pixels and "pixels with fractionary bits" - language bindings will have to choose what to bind, and resort to manually overriding the API + *except* for language bindings based on GObject-Introspection, as they cannot do manual overrides, thus will replicate the entire set of entry points For these reason, we should coalesces every Actor entry point for pixels and for ClutterUnit into a single entry point taking a float, like: void clutter_actor_set_x (ClutterActor *self, gfloat x); void clutter_actor_get_size (ClutterActor *self, gfloat *width, gfloat *height); gfloat clutter_actor_get_height (ClutterActor *self); etc. The issues I have identified are: - we'll have a two cases of compiler warnings: - printf() format of the return values from %d to %f - clutter_actor_get_size() taking floats instead of unsigned ints - we'll have a problem with varargs when passing an integer instead of a floating point value, except on 64bit platforms where the size of a float is the same as the size of an int To be clear: the *intent* of the API should not change -- we still use pixels everywhere -- but: - we remove ambiguity in the API with regard to pixels and units - we remove entry points we get to maintain for the whole 1.0 version of the API - we make things simpler to bind for both manual language bindings and automatic (gobject-introspection based) ones - we have the simplest API possible while still exposing the capabilities of the underlying GL implementation
2009-05-06 15:44:47 +00:00
clutter_anchor_coord_set_gravity (AnchorCoord *coord,
ClutterGravity gravity)
{
switch (gravity)
{
case CLUTTER_GRAVITY_NORTH:
coord->v.fraction.x = 0.5;
coord->v.fraction.y = 0.0;
break;
case CLUTTER_GRAVITY_NORTH_EAST:
coord->v.fraction.x = 1.0;
coord->v.fraction.y = 0.0;
break;
case CLUTTER_GRAVITY_EAST:
coord->v.fraction.x = 1.0;
coord->v.fraction.y = 0.5;
break;
case CLUTTER_GRAVITY_SOUTH_EAST:
coord->v.fraction.x = 1.0;
coord->v.fraction.y = 1.0;
break;
case CLUTTER_GRAVITY_SOUTH:
coord->v.fraction.x = 0.5;
coord->v.fraction.y = 1.0;
break;
case CLUTTER_GRAVITY_SOUTH_WEST:
coord->v.fraction.x = 0.0;
coord->v.fraction.y = 1.0;
break;
case CLUTTER_GRAVITY_WEST:
coord->v.fraction.x = 0.0;
coord->v.fraction.y = 0.5;
break;
case CLUTTER_GRAVITY_NORTH_WEST:
coord->v.fraction.x = 0.0;
coord->v.fraction.y = 0.0;
break;
case CLUTTER_GRAVITY_CENTER:
coord->v.fraction.x = 0.5;
coord->v.fraction.y = 0.5;
break;
default:
coord->v.fraction.x = 0.0;
coord->v.fraction.y = 0.0;
break;
}
coord->is_fractional = TRUE;
}
static gboolean
clutter_anchor_coord_is_zero (const AnchorCoord *coord)
{
if (coord->is_fractional)
return coord->v.fraction.x == 0.0 && coord->v.fraction.y == 0.0;
else
return (coord->v.units.x == 0.0
&& coord->v.units.y == 0.0
&& coord->v.units.z == 0.0);
}
/**
* clutter_actor_get_flags:
* @self: a #ClutterActor
*
* Retrieves the flags set on @self
*
* Return value: a bitwise or of #ClutterActorFlags or 0
*
* Since: 1.0
*/
ClutterActorFlags
clutter_actor_get_flags (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
return self->flags;
}
/**
* clutter_actor_set_flags:
* @self: a #ClutterActor
* @flags: the flags to set
*
* Sets @flags on @self
*
* This function will emit notifications for the changed properties
*
* Since: 1.0
*/
void
clutter_actor_set_flags (ClutterActor *self,
ClutterActorFlags flags)
{
ClutterActorFlags old_flags;
GObject *obj;
gboolean was_reactive_set, reactive_set;
gboolean was_realized_set, realized_set;
gboolean was_mapped_set, mapped_set;
gboolean was_visible_set, visible_set;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (self->flags == flags)
return;
obj = G_OBJECT (self);
g_object_ref (obj);
g_object_freeze_notify (obj);
old_flags = self->flags;
was_reactive_set = ((old_flags & CLUTTER_ACTOR_REACTIVE) != 0);
was_realized_set = ((old_flags & CLUTTER_ACTOR_REALIZED) != 0);
was_mapped_set = ((old_flags & CLUTTER_ACTOR_MAPPED) != 0);
was_visible_set = ((old_flags & CLUTTER_ACTOR_VISIBLE) != 0);
self->flags |= flags;
reactive_set = ((self->flags & CLUTTER_ACTOR_REACTIVE) != 0);
realized_set = ((self->flags & CLUTTER_ACTOR_REALIZED) != 0);
mapped_set = ((self->flags & CLUTTER_ACTOR_MAPPED) != 0);
visible_set = ((self->flags & CLUTTER_ACTOR_VISIBLE) != 0);
if (reactive_set != was_reactive_set)
g_object_notify_by_pspec (obj, obj_props[PROP_REACTIVE]);
if (realized_set != was_realized_set)
g_object_notify_by_pspec (obj, obj_props[PROP_REALIZED]);
if (mapped_set != was_mapped_set)
g_object_notify_by_pspec (obj, obj_props[PROP_MAPPED]);
if (visible_set != was_visible_set)
g_object_notify_by_pspec (obj, obj_props[PROP_VISIBLE]);
g_object_thaw_notify (obj);
g_object_unref (obj);
}
/**
* clutter_actor_unset_flags:
* @self: a #ClutterActor
* @flags: the flags to unset
*
* Unsets @flags on @self
*
* This function will emit notifications for the changed properties
*
* Since: 1.0
*/
void
clutter_actor_unset_flags (ClutterActor *self,
ClutterActorFlags flags)
{
ClutterActorFlags old_flags;
GObject *obj;
gboolean was_reactive_set, reactive_set;
gboolean was_realized_set, realized_set;
gboolean was_mapped_set, mapped_set;
gboolean was_visible_set, visible_set;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
obj = G_OBJECT (self);
g_object_freeze_notify (obj);
old_flags = self->flags;
was_reactive_set = ((old_flags & CLUTTER_ACTOR_REACTIVE) != 0);
was_realized_set = ((old_flags & CLUTTER_ACTOR_REALIZED) != 0);
was_mapped_set = ((old_flags & CLUTTER_ACTOR_MAPPED) != 0);
was_visible_set = ((old_flags & CLUTTER_ACTOR_VISIBLE) != 0);
self->flags &= ~flags;
if (self->flags == old_flags)
return;
reactive_set = ((self->flags & CLUTTER_ACTOR_REACTIVE) != 0);
realized_set = ((self->flags & CLUTTER_ACTOR_REALIZED) != 0);
mapped_set = ((self->flags & CLUTTER_ACTOR_MAPPED) != 0);
visible_set = ((self->flags & CLUTTER_ACTOR_VISIBLE) != 0);
if (reactive_set != was_reactive_set)
g_object_notify_by_pspec (obj, obj_props[PROP_REACTIVE]);
if (realized_set != was_realized_set)
g_object_notify_by_pspec (obj, obj_props[PROP_REALIZED]);
if (mapped_set != was_mapped_set)
g_object_notify_by_pspec (obj, obj_props[PROP_MAPPED]);
if (visible_set != was_visible_set)
g_object_notify_by_pspec (obj, obj_props[PROP_VISIBLE]);
g_object_thaw_notify (obj);
}
[actor] Allow changing the transformations matrix Currently, the transformation matrix for an actor is constructed from scenegraph-related accessors. An actor, though, can call COGL API to add new transformations inside the paint() implementation, for instance: static void my_foo_paint (ClutterActor *a) { ... cogl_translate (-scroll_x, -scroll_y, 0); ... } Unfortunately these transformations will be completely ignored by the scenegraph machinery; for instance, getting the actor-relative coordinates from event coordinates is going to break badly because of this. In order to make the scenegraph aware of the potential of additional transformations, we need a ::apply_transform() virtual function. This vfunc will pass a CoglMatrix which can be used to apply additional operations: static void my_foo_apply_transform (ClutterActor *a, CoglMatrix *m) { CLUTTER_ACTOR_CLASS (my_foo_parent_class)->apply_transform (a, m); ... cogl_matrix_translate (m, -scroll_x, -scroll_y, 0); ... } The ::paint() implementation will be called with the actor already using the newly applied transformation matrix, as expected: static void my_foo_paint (ClutterActor *a) { ... } The ::apply_transform() implementations *must* chain up, so that the various transformations of each class are preserved. The default implementation inside ClutterActor applies all the transformations defined by the scenegraph-related accessors. Actors performing transformations inside the paint() function will continue to work as previously.
2009-07-01 12:59:13 +00:00
/**
* clutter_actor_get_transformation_matrix:
* @self: a #ClutterActor
* @matrix: (out caller-allocates): the return location for a #CoglMatrix
[actor] Allow changing the transformations matrix Currently, the transformation matrix for an actor is constructed from scenegraph-related accessors. An actor, though, can call COGL API to add new transformations inside the paint() implementation, for instance: static void my_foo_paint (ClutterActor *a) { ... cogl_translate (-scroll_x, -scroll_y, 0); ... } Unfortunately these transformations will be completely ignored by the scenegraph machinery; for instance, getting the actor-relative coordinates from event coordinates is going to break badly because of this. In order to make the scenegraph aware of the potential of additional transformations, we need a ::apply_transform() virtual function. This vfunc will pass a CoglMatrix which can be used to apply additional operations: static void my_foo_apply_transform (ClutterActor *a, CoglMatrix *m) { CLUTTER_ACTOR_CLASS (my_foo_parent_class)->apply_transform (a, m); ... cogl_matrix_translate (m, -scroll_x, -scroll_y, 0); ... } The ::paint() implementation will be called with the actor already using the newly applied transformation matrix, as expected: static void my_foo_paint (ClutterActor *a) { ... } The ::apply_transform() implementations *must* chain up, so that the various transformations of each class are preserved. The default implementation inside ClutterActor applies all the transformations defined by the scenegraph-related accessors. Actors performing transformations inside the paint() function will continue to work as previously.
2009-07-01 12:59:13 +00:00
*
* Retrieves the transformations applied to @self relative to its
* parent.
[actor] Allow changing the transformations matrix Currently, the transformation matrix for an actor is constructed from scenegraph-related accessors. An actor, though, can call COGL API to add new transformations inside the paint() implementation, for instance: static void my_foo_paint (ClutterActor *a) { ... cogl_translate (-scroll_x, -scroll_y, 0); ... } Unfortunately these transformations will be completely ignored by the scenegraph machinery; for instance, getting the actor-relative coordinates from event coordinates is going to break badly because of this. In order to make the scenegraph aware of the potential of additional transformations, we need a ::apply_transform() virtual function. This vfunc will pass a CoglMatrix which can be used to apply additional operations: static void my_foo_apply_transform (ClutterActor *a, CoglMatrix *m) { CLUTTER_ACTOR_CLASS (my_foo_parent_class)->apply_transform (a, m); ... cogl_matrix_translate (m, -scroll_x, -scroll_y, 0); ... } The ::paint() implementation will be called with the actor already using the newly applied transformation matrix, as expected: static void my_foo_paint (ClutterActor *a) { ... } The ::apply_transform() implementations *must* chain up, so that the various transformations of each class are preserved. The default implementation inside ClutterActor applies all the transformations defined by the scenegraph-related accessors. Actors performing transformations inside the paint() function will continue to work as previously.
2009-07-01 12:59:13 +00:00
*
* Since: 1.0
*/
void
clutter_actor_get_transformation_matrix (ClutterActor *self,
CoglMatrix *matrix)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
cogl_matrix_init_identity (matrix);
_clutter_actor_apply_modelview_transform (self, matrix);
[actor] Allow changing the transformations matrix Currently, the transformation matrix for an actor is constructed from scenegraph-related accessors. An actor, though, can call COGL API to add new transformations inside the paint() implementation, for instance: static void my_foo_paint (ClutterActor *a) { ... cogl_translate (-scroll_x, -scroll_y, 0); ... } Unfortunately these transformations will be completely ignored by the scenegraph machinery; for instance, getting the actor-relative coordinates from event coordinates is going to break badly because of this. In order to make the scenegraph aware of the potential of additional transformations, we need a ::apply_transform() virtual function. This vfunc will pass a CoglMatrix which can be used to apply additional operations: static void my_foo_apply_transform (ClutterActor *a, CoglMatrix *m) { CLUTTER_ACTOR_CLASS (my_foo_parent_class)->apply_transform (a, m); ... cogl_matrix_translate (m, -scroll_x, -scroll_y, 0); ... } The ::paint() implementation will be called with the actor already using the newly applied transformation matrix, as expected: static void my_foo_paint (ClutterActor *a) { ... } The ::apply_transform() implementations *must* chain up, so that the various transformations of each class are preserved. The default implementation inside ClutterActor applies all the transformations defined by the scenegraph-related accessors. Actors performing transformations inside the paint() function will continue to work as previously.
2009-07-01 12:59:13 +00:00
}
void
_clutter_actor_set_in_clone_paint (ClutterActor *self,
gboolean is_in_clone_paint)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
self->priv->in_clone_paint = is_in_clone_paint;
}
/**
* clutter_actor_is_in_clone_paint:
* @self: a #ClutterActor
*
* Checks whether @self is being currently painted by a #ClutterClone
*
* This function is useful only inside the ::paint virtual function
* implementations or within handlers for the #ClutterActor::paint
* signal
*
* This function should not be used by applications
*
* Return value: %TRUE if the #ClutterActor is currently being painted
* by a #ClutterClone, and %FALSE otherwise
*
* Since: 1.0
*/
gboolean
clutter_actor_is_in_clone_paint (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
return self->priv->in_clone_paint;
}
static gboolean
set_direction_recursive (ClutterActor *actor,
gpointer user_data)
{
ClutterTextDirection text_dir = GPOINTER_TO_INT (user_data);
clutter_actor_set_text_direction (actor, text_dir);
return TRUE;
}
/**
* clutter_actor_set_text_direction:
* @self: a #ClutterActor
* @text_dir: the text direction for @self
*
* Sets the #ClutterTextDirection for an actor
*
* The passed text direction must not be %CLUTTER_TEXT_DIRECTION_DEFAULT
*
* If @self implements #ClutterContainer then this function will recurse
* inside all the children of @self (including the internal ones).
*
* Composite actors not implementing #ClutterContainer, or actors requiring
* special handling when the text direction changes, should connect to
* the #GObject::notify signal for the #ClutterActor:text-direction property
*
* Since: 1.2
*/
void
clutter_actor_set_text_direction (ClutterActor *self,
ClutterTextDirection text_dir)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (text_dir != CLUTTER_TEXT_DIRECTION_DEFAULT);
priv = self->priv;
if (priv->text_direction != text_dir)
{
priv->text_direction = text_dir;
/* we need to emit the notify::text-direction first, so that
* the sub-classes can catch that and do specific handling of
* the text direction; see clutter_text_direction_changed_cb()
* inside clutter-text.c
*/
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_TEXT_DIRECTION]);
_clutter_actor_foreach_child (self, set_direction_recursive,
GINT_TO_POINTER (text_dir));
clutter_actor_queue_relayout (self);
}
}
void
_clutter_actor_set_has_pointer (ClutterActor *self,
gboolean has_pointer)
{
ClutterActorPrivate *priv = self->priv;
if (priv->has_pointer != has_pointer)
{
priv->has_pointer = has_pointer;
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_HAS_POINTER]);
}
}
/**
* clutter_actor_get_text_direction:
* @self: a #ClutterActor
*
* Retrieves the value set using clutter_actor_set_text_direction()
*
* If no text direction has been previously set, the default text
* direction, as returned by clutter_get_default_text_direction(), will
* be returned instead
*
* Return value: the #ClutterTextDirection for the actor
*
* Since: 1.2
*/
ClutterTextDirection
clutter_actor_get_text_direction (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self),
CLUTTER_TEXT_DIRECTION_LTR);
priv = self->priv;
/* if no direction has been set yet use the default */
if (priv->text_direction == CLUTTER_TEXT_DIRECTION_DEFAULT)
priv->text_direction = clutter_get_default_text_direction ();
return priv->text_direction;
}
/**
* clutter_actor_push_internal:
* @self: a #ClutterActor
*
* Should be used by actors implementing the #ClutterContainer and with
* internal children added through clutter_actor_set_parent(), for instance:
*
* |[
* static void
* my_actor_init (MyActor *self)
* {
* self->priv = SELF_ACTOR_GET_PRIVATE (self);
*
* clutter_actor_push_internal (CLUTTER_ACTOR (self));
*
* /&ast; calling clutter_actor_set_parent() now will result in
* &ast; the internal flag being set on a child of MyActor
* &ast;/
*
* /&ast; internal child - a background texture &ast;/
* self->priv->background_tex = clutter_texture_new ();
* clutter_actor_set_parent (self->priv->background_tex,
* CLUTTER_ACTOR (self));
*
* /&ast; internal child - a label &ast;/
* self->priv->label = clutter_text_new ();
* clutter_actor_set_parent (self->priv->label,
* CLUTTER_ACTOR (self));
*
* clutter_actor_pop_internal (CLUTTER_ACTOR (self));
*
* /&ast; calling clutter_actor_set_parent() now will not result in
* &ast; the internal flag being set on a child of MyActor
* &ast;/
* }
* ]|
*
* This function will be used by Clutter to toggle an "internal child"
* flag whenever clutter_actor_set_parent() is called; internal children
* are handled differently by Clutter, specifically when destroying their
* parent.
*
* Call clutter_actor_pop_internal() when you finished adding internal
* children.
*
* Nested calls to clutter_actor_push_internal() are allowed, but each
* one must by followed by a clutter_actor_pop_internal() call.
*
* Since: 1.2
*
* Deprecated: 1.10: All children of an actor are accessible through
* the #ClutterActor API, and #ClutterActor implements the
* #ClutterContainer interface, so this function is only useful
* for legacy containers overriding the default implementation.
*/
void
clutter_actor_push_internal (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
self->priv->internal_child += 1;
}
/**
* clutter_actor_pop_internal:
* @self: a #ClutterActor
*
* Disables the effects of clutter_actor_push_internal().
*
* Since: 1.2
*
* Deprecated: 1.10: All children of an actor are accessible through
* the #ClutterActor API. This function is only useful for legacy
* containers overriding the default implementation of the
* #ClutterContainer interface.
*/
void
clutter_actor_pop_internal (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
if (priv->internal_child == 0)
{
g_warning ("Mismatched %s: you need to call "
"clutter_actor_push_composite() at least once before "
"calling this function", G_STRFUNC);
return;
}
priv->internal_child -= 1;
}
/**
* clutter_actor_has_pointer:
* @self: a #ClutterActor
*
* Checks whether an actor contains the pointer of a
* #ClutterInputDevice
*
* Return value: %TRUE if the actor contains the pointer, and
* %FALSE otherwise
*
* Since: 1.2
*/
gboolean
clutter_actor_has_pointer (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
return self->priv->has_pointer;
}
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
/* XXX: This is a workaround for not being able to break the ABI of
* the QUEUE_REDRAW signal. It is an out-of-band argument. See
* clutter_actor_queue_clipped_redraw() for details.
*/
ClutterPaintVolume *
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
_clutter_actor_get_queue_redraw_clip (ClutterActor *self)
{
return g_object_get_data (G_OBJECT (self),
"-clutter-actor-queue-redraw-clip");
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
}
void
_clutter_actor_set_queue_redraw_clip (ClutterActor *self,
ClutterPaintVolume *clip)
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
{
g_object_set_data (G_OBJECT (self),
"-clutter-actor-queue-redraw-clip",
clip);
Adds initial clipped redraw support to Clutter A new (internal only currently) API, _clutter_actor_queue_clipped_redraw can be used to queue a redraw along with a clip rectangle in actor coordinates. This clip rectangle propagates up to the stage and clutter backend which may optionally use the information to optimize stage redraws. The GLX backend in particular may scissor the next redraw to the clip rectangle and use GLX_MESA_copy_sub_buffer to present the stage subregion. The intention is that any actors that can naturally determine the bounds of updates should queue clipped redraws to reduce the cost of updating small regions of the screen. Notes: » If GLX_MESA_copy_sub_buffer isn't available then the GLX backend ignores any clip rectangles. » queuing multiple clipped redraws will result in the bounding box of each clip rectangle being used. » If a clipped redraw has a height > 300 pixels then it's promoted into a full stage redraw, so that the GPU doesn't end up blocking too long waiting for the vsync to reach the optimal position to avoid tearing. » Note: no empirical data was used to come up with this threshold so we may need to tune this. » Currently only ClutterX11TexturePixmap makes use of this new API. This is done via a new "queue-damage-redraw" signal that is emitted when the pixmap is updated. The default handler queues a clipped redraw with the assumption that the pixmap is being painted as a rectangle covering the actors transformed allocation. If you subclass ClutterX11TexturePixmap and change how it's painted you now also need to override the signal handler and queue your own redraw. Technically this is a semantic break, but it's assumed that no one is currently doing this. This still leaves a few unsolved issues with regards to optimizing sub stage redraws that need to be addressed in further work so this can only be considered a stepping stone a this point: » Because we have no reliable way to determine if the painting of any given actor is being modified any optimizations implemented using _clutter_actor_queue_redraw_with_clip must be overridable by a subclass, and technically must be opt-in for existing classes to avoid a change in semantics. E.g. consider that a user connects to the paint signal for ClutterTexture and paints a circle instead of a rectangle. In this case any original logic to queue clipped redraws would be incorrect. » Currently only the implementation of an actor has enough information with which to queue clipped redraws. E.g. It is not possible for generic code in clutter-actor.c to queue a clipped redraw when hiding an actor because actors have no way to report a "paint box". (remember actors can draw outside their allocation and actors with depth may also be projected outside of their allocation) » The current plan is to add a actor_class->get_paint_cuboid() virtual so actors can report a bounding cube for everything they would draw in their current state and use that to queue clipped redraws against the stage by projecting the paint cube into stage coordinates. » Our heuristics for promoting clipped redraws into full redraws to avoid blocking the GPU while we wait for the vsync need improving: » vsync issues aren't relevant for redirected/composited applications so they should use different heuristics. In this case we instead need to trade off the cost of blitting when using glXCopySubBuffer vs promoting to a full redraw and flipping instead.
2009-11-30 17:47:55 +00:00
}
/**
* clutter_actor_has_allocation:
* @self: a #ClutterActor
*
* Checks if the actor has an up-to-date allocation assigned to
* it. This means that the actor should have an allocation: it's
* visible and has a parent. It also means that there is no
* outstanding relayout request in progress for the actor or its
* children (There might be other outstanding layout requests in
* progress that will cause the actor to get a new allocation
* when the stage is laid out, however).
*
* If this function returns %FALSE, then the actor will normally
* be allocated before it is next drawn on the screen.
*
* Return value: %TRUE if the actor has an up-to-date allocation
*
* Since: 1.4
*/
gboolean
clutter_actor_has_allocation (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
priv = self->priv;
return priv->parent != NULL &&
CLUTTER_ACTOR_IS_VISIBLE (self) &&
!priv->needs_allocation;
}
/**
* clutter_actor_add_action:
* @self: a #ClutterActor
* @action: a #ClutterAction
*
* Adds @action to the list of actions applied to @self
*
* A #ClutterAction can only belong to one actor at a time
*
* The #ClutterActor will hold a reference on @action until either
* clutter_actor_remove_action() or clutter_actor_clear_actions()
* is called
*
* Since: 1.4
*/
void
clutter_actor_add_action (ClutterActor *self,
ClutterAction *action)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTION (action));
priv = self->priv;
if (priv->actions == NULL)
{
priv->actions = g_object_new (CLUTTER_TYPE_META_GROUP, NULL);
priv->actions->actor = self;
}
_clutter_meta_group_add_meta (priv->actions, CLUTTER_ACTOR_META (action));
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_ACTIONS]);
}
/**
* clutter_actor_add_action_with_name:
* @self: a #ClutterActor
* @name: the name to set on the action
2010-05-20 10:45:25 +00:00
* @action: a #ClutterAction
*
* A convenience function for setting the name of a #ClutterAction
* while adding it to the list of actions applied to @self
*
* This function is the logical equivalent of:
*
* |[
* clutter_actor_meta_set_name (CLUTTER_ACTOR_META (action), name);
* clutter_actor_add_action (self, action);
* ]|
*
* Since: 1.4
*/
void
clutter_actor_add_action_with_name (ClutterActor *self,
const gchar *name,
ClutterAction *action)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (name != NULL);
g_return_if_fail (CLUTTER_IS_ACTION (action));
clutter_actor_meta_set_name (CLUTTER_ACTOR_META (action), name);
clutter_actor_add_action (self, action);
}
/**
* clutter_actor_remove_action:
* @self: a #ClutterActor
* @action: a #ClutterAction
*
* Removes @action from the list of actions applied to @self
*
* The reference held by @self on the #ClutterAction will be released
*
* Since: 1.4
*/
void
clutter_actor_remove_action (ClutterActor *self,
ClutterAction *action)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_ACTION (action));
priv = self->priv;
if (priv->actions == NULL)
return;
_clutter_meta_group_remove_meta (priv->actions, CLUTTER_ACTOR_META (action));
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_ACTIONS]);
}
/**
* clutter_actor_remove_action_by_name:
* @self: a #ClutterActor
* @name: the name of the action to remove
*
* Removes the #ClutterAction with the given name from the list
* of actions applied to @self
*
* Since: 1.4
*/
void
clutter_actor_remove_action_by_name (ClutterActor *self,
const gchar *name)
{
ClutterActorPrivate *priv;
ClutterActorMeta *meta;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (name != NULL);
priv = self->priv;
if (priv->actions == NULL)
return;
meta = _clutter_meta_group_get_meta (priv->actions, name);
if (meta == NULL)
return;
_clutter_meta_group_remove_meta (priv->actions, meta);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_ACTIONS]);
}
/**
* clutter_actor_get_actions:
* @self: a #ClutterActor
*
* Retrieves the list of actions applied to @self
*
2010-09-08 15:41:47 +00:00
* Return value: (transfer container) (element-type Clutter.Action): a copy
* of the list of #ClutterAction<!-- -->s. The contents of the list are
* owned by the #ClutterActor. Use g_list_free() to free the resources
* allocated by the returned #GList
*
* Since: 1.4
*/
GList *
clutter_actor_get_actions (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
if (self->priv->actions == NULL)
return NULL;
return _clutter_meta_group_get_metas_no_internal (self->priv->actions);
}
/**
* clutter_actor_get_action:
* @self: a #ClutterActor
* @name: the name of the action to retrieve
*
* Retrieves the #ClutterAction with the given name in the list
* of actions applied to @self
*
* Return value: (transfer none): a #ClutterAction for the given
* name, or %NULL. The returned #ClutterAction is owned by the
* actor and it should not be unreferenced directly
*
* Since: 1.4
*/
ClutterAction *
clutter_actor_get_action (ClutterActor *self,
const gchar *name)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
g_return_val_if_fail (name != NULL, NULL);
if (self->priv->actions == NULL)
return NULL;
return CLUTTER_ACTION (_clutter_meta_group_get_meta (self->priv->actions, name));
}
/**
* clutter_actor_clear_actions:
* @self: a #ClutterActor
*
* Clears the list of actions applied to @self
*
* Since: 1.4
*/
void
clutter_actor_clear_actions (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (self->priv->actions == NULL)
return;
_clutter_meta_group_clear_metas_no_internal (self->priv->actions);
}
/**
* clutter_actor_add_constraint:
* @self: a #ClutterActor
* @constraint: a #ClutterConstraint
*
* Adds @constraint to the list of #ClutterConstraint<!-- -->s applied
* to @self
*
* The #ClutterActor will hold a reference on the @constraint until
* either clutter_actor_remove_constraint() or
* clutter_actor_clear_constraints() is called.
*
* Since: 1.4
*/
void
clutter_actor_add_constraint (ClutterActor *self,
ClutterConstraint *constraint)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_CONSTRAINT (constraint));
priv = self->priv;
if (priv->constraints == NULL)
{
priv->constraints = g_object_new (CLUTTER_TYPE_META_GROUP, NULL);
priv->constraints->actor = self;
}
_clutter_meta_group_add_meta (priv->constraints,
CLUTTER_ACTOR_META (constraint));
clutter_actor_queue_relayout (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_CONSTRAINTS]);
}
/**
* clutter_actor_add_constraint_with_name:
* @self: a #ClutterActor
* @name: the name to set on the constraint
* @constraint: a #ClutterConstraint
*
* A convenience function for setting the name of a #ClutterConstraint
* while adding it to the list of constraints applied to @self
*
* This function is the logical equivalent of:
*
* |[
* clutter_actor_meta_set_name (CLUTTER_ACTOR_META (constraint), name);
* clutter_actor_add_constraint (self, constraint);
* ]|
*
* Since: 1.4
*/
void
clutter_actor_add_constraint_with_name (ClutterActor *self,
const gchar *name,
ClutterConstraint *constraint)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (name != NULL);
g_return_if_fail (CLUTTER_IS_CONSTRAINT (constraint));
clutter_actor_meta_set_name (CLUTTER_ACTOR_META (constraint), name);
clutter_actor_add_constraint (self, constraint);
}
/**
* clutter_actor_remove_constraint:
* @self: a #ClutterActor
* @constraint: a #ClutterConstraint
*
* Removes @constraint from the list of constraints applied to @self
*
* The reference held by @self on the #ClutterConstraint will be released
*
* Since: 1.4
*/
void
clutter_actor_remove_constraint (ClutterActor *self,
ClutterConstraint *constraint)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_CONSTRAINT (constraint));
priv = self->priv;
if (priv->constraints == NULL)
return;
_clutter_meta_group_remove_meta (priv->constraints,
CLUTTER_ACTOR_META (constraint));
clutter_actor_queue_relayout (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_CONSTRAINTS]);
}
/**
* clutter_actor_remove_constraint_by_name:
* @self: a #ClutterActor
* @name: the name of the constraint to remove
*
* Removes the #ClutterConstraint with the given name from the list
* of constraints applied to @self
*
* Since: 1.4
*/
void
clutter_actor_remove_constraint_by_name (ClutterActor *self,
const gchar *name)
{
ClutterActorPrivate *priv;
ClutterActorMeta *meta;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (name != NULL);
priv = self->priv;
if (priv->constraints == NULL)
return;
meta = _clutter_meta_group_get_meta (priv->constraints, name);
if (meta == NULL)
return;
_clutter_meta_group_remove_meta (priv->constraints, meta);
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_get_constraints:
* @self: a #ClutterActor
*
* Retrieves the list of constraints applied to @self
*
2010-09-08 15:41:47 +00:00
* Return value: (transfer container) (element-type Clutter.Constraint): a copy
* of the list of #ClutterConstraint<!-- -->s. The contents of the list are
* owned by the #ClutterActor. Use g_list_free() to free the resources
* allocated by the returned #GList
*
* Since: 1.4
*/
GList *
clutter_actor_get_constraints (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
if (self->priv->constraints == NULL)
return NULL;
return _clutter_meta_group_get_metas_no_internal (self->priv->constraints);
}
/**
* clutter_actor_get_constraint:
* @self: a #ClutterActor
* @name: the name of the constraint to retrieve
*
* Retrieves the #ClutterConstraint with the given name in the list
* of constraints applied to @self
*
* Return value: (transfer none): a #ClutterConstraint for the given
* name, or %NULL. The returned #ClutterConstraint is owned by the
* actor and it should not be unreferenced directly
*
* Since: 1.4
*/
ClutterConstraint *
clutter_actor_get_constraint (ClutterActor *self,
const gchar *name)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
g_return_val_if_fail (name != NULL, NULL);
if (self->priv->constraints == NULL)
return NULL;
return CLUTTER_CONSTRAINT (_clutter_meta_group_get_meta (self->priv->constraints, name));
}
/**
* clutter_actor_clear_constraints:
* @self: a #ClutterActor
*
* Clears the list of constraints applied to @self
*
* Since: 1.4
*/
void
clutter_actor_clear_constraints (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (self->priv->constraints == NULL)
return;
_clutter_meta_group_clear_metas_no_internal (self->priv->constraints);
clutter_actor_queue_relayout (self);
}
/**
* clutter_actor_set_clip_to_allocation:
* @self: a #ClutterActor
* @clip_set: %TRUE to apply a clip tracking the allocation
*
* Sets whether @self should be clipped to the same size as its
* allocation
*
* Since: 1.4
*/
void
clutter_actor_set_clip_to_allocation (ClutterActor *self,
gboolean clip_set)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
clip_set = !!clip_set;
priv = self->priv;
if (priv->clip_to_allocation != clip_set)
{
priv->clip_to_allocation = clip_set;
clutter_actor_queue_redraw (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_CLIP_TO_ALLOCATION]);
}
}
/**
* clutter_actor_get_clip_to_allocation:
* @self: a #ClutterActor
*
* Retrieves the value set using clutter_actor_set_clip_to_allocation()
*
* Return value: %TRUE if the #ClutterActor is clipped to its allocation
*
* Since: 1.4
*/
gboolean
clutter_actor_get_clip_to_allocation (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
return self->priv->clip_to_allocation;
}
/**
* clutter_actor_add_effect:
* @self: a #ClutterActor
* @effect: a #ClutterEffect
*
* Adds @effect to the list of #ClutterEffect<!-- -->s applied to @self
*
* The #ClutterActor will hold a reference on the @effect until either
* clutter_actor_remove_effect() or clutter_actor_clear_effects() is
* called.
*
* Since: 1.4
*/
void
clutter_actor_add_effect (ClutterActor *self,
ClutterEffect *effect)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_EFFECT (effect));
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
_clutter_actor_add_effect_internal (self, effect);
clutter_actor_queue_redraw (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_EFFECT]);
}
/**
* clutter_actor_add_effect_with_name:
* @self: a #ClutterActor
* @name: the name to set on the effect
* @effect: a #ClutterEffect
*
* A convenience function for setting the name of a #ClutterEffect
* while adding it to the list of effectss applied to @self
*
* This function is the logical equivalent of:
*
* |[
* clutter_actor_meta_set_name (CLUTTER_ACTOR_META (effect), name);
* clutter_actor_add_effect (self, effect);
* ]|
*
* Since: 1.4
*/
void
clutter_actor_add_effect_with_name (ClutterActor *self,
const gchar *name,
ClutterEffect *effect)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (name != NULL);
g_return_if_fail (CLUTTER_IS_EFFECT (effect));
clutter_actor_meta_set_name (CLUTTER_ACTOR_META (effect), name);
clutter_actor_add_effect (self, effect);
}
/**
* clutter_actor_remove_effect:
* @self: a #ClutterActor
* @effect: a #ClutterEffect
*
* Removes @effect from the list of effects applied to @self
*
* The reference held by @self on the #ClutterEffect will be released
*
* Since: 1.4
*/
void
clutter_actor_remove_effect (ClutterActor *self,
ClutterEffect *effect)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (CLUTTER_IS_EFFECT (effect));
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
_clutter_actor_remove_effect_internal (self, effect);
clutter_actor_queue_redraw (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_EFFECT]);
}
/**
* clutter_actor_remove_effect_by_name:
* @self: a #ClutterActor
* @name: the name of the effect to remove
*
* Removes the #ClutterEffect with the given name from the list
* of effects applied to @self
*
* Since: 1.4
*/
void
clutter_actor_remove_effect_by_name (ClutterActor *self,
const gchar *name)
{
ClutterActorPrivate *priv;
ClutterActorMeta *meta;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (name != NULL);
priv = self->priv;
if (priv->effects == NULL)
return;
meta = _clutter_meta_group_get_meta (priv->effects, name);
if (meta == NULL)
return;
clutter_actor_remove_effect (self, CLUTTER_EFFECT (meta));
}
/**
* clutter_actor_get_effects:
* @self: a #ClutterActor
*
* Retrieves the #ClutterEffect<!-- -->s applied on @self, if any
*
2010-09-08 15:41:47 +00:00
* Return value: (transfer container) (element-type Clutter.Effect): a list
* of #ClutterEffect<!-- -->s, or %NULL. The elements of the returned
* list are owned by Clutter and they should not be freed. You should
* free the returned list using g_list_free() when done
*
* Since: 1.4
*/
GList *
clutter_actor_get_effects (ClutterActor *self)
{
ClutterActorPrivate *priv;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
priv = self->priv;
if (priv->effects == NULL)
return NULL;
return _clutter_meta_group_get_metas_no_internal (priv->effects);
}
/**
* clutter_actor_get_effect:
* @self: a #ClutterActor
* @name: the name of the effect to retrieve
*
* Retrieves the #ClutterEffect with the given name in the list
* of effects applied to @self
*
* Return value: (transfer none): a #ClutterEffect for the given
* name, or %NULL. The returned #ClutterEffect is owned by the
* actor and it should not be unreferenced directly
*
* Since: 1.4
*/
ClutterEffect *
clutter_actor_get_effect (ClutterActor *self,
const gchar *name)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
g_return_val_if_fail (name != NULL, NULL);
if (self->priv->effects == NULL)
return NULL;
return CLUTTER_EFFECT (_clutter_meta_group_get_meta (self->priv->effects, name));
}
/**
* clutter_actor_clear_effects:
* @self: a #ClutterActor
*
* Clears the list of effects applied to @self
*
* Since: 1.4
*/
void
clutter_actor_clear_effects (ClutterActor *self)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
if (self->priv->effects == NULL)
return;
_clutter_meta_group_clear_metas_no_internal (self->priv->effects);
clutter_actor_queue_redraw (self);
}
/**
* clutter_actor_has_key_focus:
* @self: a #ClutterActor
*
* Checks whether @self is the #ClutterActor that has key focus
*
* Return value: %TRUE if the actor has key focus, and %FALSE otherwise
*
* Since: 1.4
*/
gboolean
clutter_actor_has_key_focus (ClutterActor *self)
{
ClutterActor *stage;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
stage = _clutter_actor_get_stage_internal (self);
if (stage == NULL)
return FALSE;
return clutter_stage_get_key_focus (CLUTTER_STAGE (stage)) == self;
}
static gboolean
_clutter_actor_get_paint_volume_real (ClutterActor *self,
ClutterPaintVolume *pv)
{
ClutterActorPrivate *priv = self->priv;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
/* Actors are only expected to report a valid paint volume
* while they have a valid allocation. */
if (G_UNLIKELY (priv->needs_allocation))
{
CLUTTER_NOTE (CLIPPING, "Bail from get_paint_volume (%s): "
"Actor needs allocation",
_clutter_actor_get_debug_name (self));
return FALSE;
}
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
/* Check if there are any handlers connected to the paint
* signal. If there are then all bets are off for what the paint
* volume for this actor might possibly be!
*
* XXX: It's expected that this is going to end up being quite a
* costly check to have to do here, but we haven't come up with
* another solution that can reliably catch paint signal handlers at
* the right time to either avoid artefacts due to invalid stage
* clipping or due to incorrect culling.
*
* Previously we checked in clutter_actor_paint(), but at that time
* we may already be using a stage clip that could be derived from
* an invalid paint-volume. We used to try and handle that by
* queuing a follow up, unclipped, redraw but still the previous
* checking wasn't enough to catch invalid volumes involved in
* culling (considering that containers may derive their volume from
* children that haven't yet been painted)
*
* Longer term, improved solutions could be:
* - Disallow painting in the paint signal, only allow using it
* for tracking when paints happen. We can add another API that
* allows monkey patching the paint of arbitrary actors but in a
* more controlled way and that also supports modifying the
* paint-volume.
* - If we could be notified somehow when signal handlers are
* connected we wouldn't have to poll for handlers like this.
*/
if (g_signal_has_handler_pending (self,
actor_signals[PAINT],
0,
TRUE))
{
CLUTTER_NOTE (CLIPPING, "Bail from get_paint_volume (%s): "
"Actor has \"paint\" signal handlers",
_clutter_actor_get_debug_name (self));
return FALSE;
}
_clutter_paint_volume_init_static (pv, self);
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
if (!CLUTTER_ACTOR_GET_CLASS (self)->get_paint_volume (self, pv))
{
clutter_paint_volume_free (pv);
CLUTTER_NOTE (CLIPPING, "Bail from get_paint_volume (%s): "
"Actor failed to report a volume",
_clutter_actor_get_debug_name (self));
return FALSE;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
}
/* since effects can modify the paint volume, we allow them to actually
* do this by making get_paint_volume() "context sensitive"
*/
if (priv->effects != NULL)
{
if (priv->current_effect != NULL)
{
const GList *effects, *l;
/* if we are being called from within the paint sequence of
* an actor, get the paint volume up to the current effect
*/
effects = _clutter_meta_group_peek_metas (priv->effects);
for (l = effects;
l != NULL || (l != NULL && l->data != priv->current_effect);
l = l->next)
{
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
if (!_clutter_effect_get_paint_volume (l->data, pv))
{
clutter_paint_volume_free (pv);
CLUTTER_NOTE (CLIPPING, "Bail from get_paint_volume (%s): "
"Effect (%s) failed to report a volume",
_clutter_actor_get_debug_name (self),
_clutter_actor_meta_get_debug_name (l->data));
return FALSE;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
}
}
}
else
{
const GList *effects, *l;
/* otherwise, get the cumulative volume */
effects = _clutter_meta_group_peek_metas (priv->effects);
for (l = effects; l != NULL; l = l->next)
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
if (!_clutter_effect_get_paint_volume (l->data, pv))
{
clutter_paint_volume_free (pv);
CLUTTER_NOTE (CLIPPING, "Bail from get_paint_volume (%s): "
"Effect (%s) failed to report a volume",
_clutter_actor_get_debug_name (self),
_clutter_actor_meta_get_debug_name (l->data));
return FALSE;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
}
}
}
return TRUE;
}
/* The public clutter_actor_get_paint_volume API returns a const
* pointer since we return a pointer directly to the cached
* PaintVolume associated with the actor and don't want the user to
* inadvertently modify it, but for internal uses we sometimes need
* access to the same PaintVolume but need to apply some book-keeping
* modifications to it so we don't want a const pointer.
*/
static ClutterPaintVolume *
_clutter_actor_get_paint_volume_mutable (ClutterActor *self)
{
ClutterActorPrivate *priv;
priv = self->priv;
if (priv->paint_volume_valid)
clutter_paint_volume_free (&priv->paint_volume);
if (_clutter_actor_get_paint_volume_real (self, &priv->paint_volume))
{
priv->paint_volume_valid = TRUE;
return &priv->paint_volume;
}
else
{
priv->paint_volume_valid = FALSE;
return NULL;
}
}
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
/**
* clutter_actor_get_paint_volume:
* @self: a #ClutterActor
*
* Retrieves the paint volume of the passed #ClutterActor, or %NULL
* when a paint volume can't be determined.
*
* The paint volume is defined as the 3D space occupied by an actor
* when being painted.
*
* This function will call the <function>get_paint_volume()</function>
* virtual function of the #ClutterActor class. Sub-classes of #ClutterActor
* should not usually care about overriding the default implementation,
* unless they are, for instance: painting outside their allocation, or
* actors with a depth factor (not in terms of #ClutterActor:depth but real
* 3D depth).
*
* <note>2D actors overriding <function>get_paint_volume()</function>
* ensure their volume has a depth of 0. (This will be true so long as
* you don't call clutter_paint_volume_set_depth().)</note>
*
* Return value: (transfer none): a pointer to a #ClutterPaintVolume
* or %NULL if no volume could be determined.
*
* Since: 1.6
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
*/
const ClutterPaintVolume *
clutter_actor_get_paint_volume (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
return _clutter_actor_get_paint_volume_mutable (self);
}
/**
* clutter_actor_get_transformed_paint_volume:
* @self: a #ClutterActor
* @relative_to_ancestor: A #ClutterActor that is an ancestor of @self
* (or %NULL for the stage)
*
* Retrieves the 3D paint volume of an actor like
* clutter_actor_get_paint_volume() does (Please refer to the
* documentation of clutter_actor_get_paint_volume() for more
* details.) and it additionally transforms the paint volume into the
* coordinate space of @relative_to_ancestor. (Or the stage if %NULL
* is passed for @relative_to_ancestor)
*
* This can be used by containers that base their paint volume on
* the volume of their children. Such containers can query the
* transformed paint volume of all of its children and union them
* together using clutter_paint_volume_union().
*
* Return value: (transfer none): a pointer to a #ClutterPaintVolume
* or %NULL if no volume could be determined.
*
* Since: 1.6
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
*/
const ClutterPaintVolume *
clutter_actor_get_transformed_paint_volume (ClutterActor *self,
ClutterActor *relative_to_ancestor)
{
const ClutterPaintVolume *volume;
ClutterActor *stage;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
ClutterPaintVolume *transformed_volume;
stage = _clutter_actor_get_stage_internal (self);
if (G_UNLIKELY (stage == NULL))
return NULL;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
if (relative_to_ancestor == NULL)
relative_to_ancestor = stage;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
volume = clutter_actor_get_paint_volume (self);
if (volume == NULL)
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
return NULL;
transformed_volume =
_clutter_stage_paint_volume_stack_allocate (CLUTTER_STAGE (stage));
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
_clutter_paint_volume_copy_static (volume, transformed_volume);
_clutter_paint_volume_transform_relative (transformed_volume,
relative_to_ancestor);
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
return transformed_volume;
}
/**
* clutter_actor_get_paint_box:
* @self: a #ClutterActor
* @box: (out): return location for a #ClutterActorBox
*
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
* Retrieves the paint volume of the passed #ClutterActor, and
* transforms it into a 2D bounding box in stage coordinates.
*
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
* This function is useful to determine the on screen area occupied by
* the actor. The box is only an approximation and may often be
* considerably larger due to the optimizations used to calculate the
* box. The box is never smaller though, so it can reliably be used
* for culling.
*
* There are times when a 2D paint box can't be determined, e.g.
* because the actor isn't yet parented under a stage or because
* the actor is unable to determine a paint volume.
*
* Return value: %TRUE if a 2D paint box could be determined, else
* %FALSE.
*
* Since: 1.6
*/
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
gboolean
clutter_actor_get_paint_box (ClutterActor *self,
ClutterActorBox *box)
{
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
ClutterActor *stage;
ClutterPaintVolume *pv;
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), FALSE);
g_return_val_if_fail (box != NULL, FALSE);
stage = _clutter_actor_get_stage_internal (self);
if (G_UNLIKELY (!stage))
return FALSE;
pv = _clutter_actor_get_paint_volume_mutable (self);
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
if (G_UNLIKELY (!pv))
return FALSE;
_clutter_paint_volume_get_stage_paint_box (pv, CLUTTER_STAGE (stage), box);
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
return TRUE;
}
paint volumes: another pass at the design This is a fairly extensive second pass at exposing paint volumes for actors. The API has changed to allow clutter_actor_get_paint_volume to fail since there are times - such as when an actor isn't a descendent of the stage - when the volume can't be determined. Another example is when something has connected to the "paint" signal of the actor and we simply have no way of knowing what might be drawn in that handler. The API has also be changed to return a const ClutterPaintVolume pointer (transfer none) so we can avoid having to dynamically allocate the volumes in the most common/performance critical code paths. Profiling was showing the slice allocation of volumes taking about 1% of an apps time, for some fairly basic tests. Most volumes can now simply be allocated on the stack; for clutter_actor_get_paint_volume we return a pointer to &priv->paint_volume and if we need a more dynamic allocation there is now a _clutter_stage_paint_volume_stack_allocate() mechanism which lets us allocate data which expires at the start of the next frame. The API has been extended to make it easier to implement get_paint_volume for containers by using clutter_actor_get_transformed_paint_volume and clutter_paint_volume_union. The first allows you to query the paint volume of a child but transformed into parent actor coordinates. The second lets you combine volumes together so you can union all the volumes for a container's children and report that as the container's own volume. The representation of paint volumes has been updated to consider that 2D actors are the most common. The effect apis, clutter-texture and clutter-group have been update accordingly.
2010-09-07 17:04:19 +00:00
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
/**
* clutter_actor_has_overlaps:
* @self: A #ClutterActor
*
2011-06-13 12:27:46 +00:00
* Asks the actor's implementation whether it may contain overlapping
* primitives.
*
* For example; Clutter may use this to determine whether the painting
* should be redirected to an offscreen buffer to correctly implement
* the opacity property.
2011-06-13 12:27:46 +00:00
*
* Custom actors can override the default response by implementing the
* #ClutterActor <function>has_overlaps</function> virtual function. See
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
* clutter_actor_set_offscreen_redirect() for more information.
*
2011-06-13 12:27:46 +00:00
* Return value: %TRUE if the actor may have overlapping primitives, and
* %FALSE otherwise
*
clutter-actor: Add a 'has_overlaps' virtual This adds a virtual to ClutterActor so that an actor subclass can report whether it has overlapping primitives. ClutterActor uses this to determine whether it needs to use ClutterFlattenEffect to implement the opacity property. The default implementation of the virtual returns TRUE which means that most actors will end up being redirected offscreen when the opacity != 255. ClutterTexture and ClutterRectangle override this to return FALSE because they should never need to be redirected. ClutterClone overrides it to divert to the source. The values for the ClutterOffscreenRedirect enum have changed to: AUTOMATIC_FOR_OPACITY The actor will only be redirected if has_overlaps returns TRUE and the opacity is < 255 ALWAYS_FOR_OPACITY The actor will always be redirected if the opacity < 255 regardless of the return value of has_overlaps ALWAYS The actor will always be redirected offscreen. This means that the property can't be used to prevent the actor from being redirected but only to increase the likelihood that it will be redirected. ClutterActor now adds and removes the flatten effect depending on whether flattening is needed directly in clutter_actor_paint(). There are new internal versions of add/remove_effect that don't queue a redraw. This means that ClutterFlattenEffect is now just a no-op subclass of ClutterOffscreen. It is only needed because ClutterOffscreen is abstract. Removing the effect also makes it so that the cached image will be freed as soon as an actor is repainted without being flattened.
2011-03-14 16:30:53 +00:00
* Since: 1.8
*/
gboolean
clutter_actor_has_overlaps (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), TRUE);
return CLUTTER_ACTOR_GET_CLASS (self)->has_overlaps (self);
}
/**
* clutter_actor_has_effects:
* @self: A #ClutterActor
*
* Returns whether the actor has any effects applied.
*
* Return value: %TRUE if the actor has any effects,
* %FALSE otherwise
*
* Since: 1.10
*/
gboolean
clutter_actor_has_effects (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), TRUE);
if (self->priv->effects == NULL)
return FALSE;
return _clutter_meta_group_has_metas_no_internal (self->priv->effects);
}
/**
* clutter_actor_has_constraints:
* @self: A #ClutterActor
*
* Returns whether the actor has any constraints applied.
*
* Return value: %TRUE if the actor has any constraints,
* %FALSE otherwise
*
* Since: 1.10
*/
gboolean
clutter_actor_has_constraints (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), TRUE);
return self->priv->constraints != NULL;
}
/**
* clutter_actor_has_actions:
* @self: A #ClutterActor
*
* Returns whether the actor has any actions applied.
*
* Return value: %TRUE if the actor has any actions,
* %FALSE otherwise
*
* Since: 1.10
*/
gboolean
clutter_actor_has_actions (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), TRUE);
return self->priv->actions != NULL;
}
/**
* clutter_actor_get_n_children:
* @self: a #ClutterActor
*
* Retrieves the number of children of @self.
*
* Return value: the number of children of an actor
*
* Since: 1.10
*/
gint
clutter_actor_get_n_children (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0);
return self->priv->n_children;
}
/**
* clutter_actor_get_child_at_index:
* @self: a #ClutterActor
* @index_: the position in the list of children
*
* Retrieves the actor at the given @index_ inside the list of
* children of @self.
*
* Return value: (transfer none): a pointer to a #ClutterActor, or %NULL
*
* Since: 1.10
*/
ClutterActor *
clutter_actor_get_child_at_index (ClutterActor *self,
gint index_)
{
ClutterActor *iter;
int i;
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
g_return_val_if_fail (index_ <= self->priv->n_children, NULL);
for (iter = self->priv->first_child, i = 0;
iter != NULL && i < index_;
iter = iter->priv->next_sibling, i += 1)
;
return iter;
}
/*< private >
* _clutter_actor_foreach_child:
* @actor: The actor whos children you want to iterate
* @callback: The function to call for each child
* @user_data: Private data to pass to @callback
*
* Calls a given @callback once for each child of the specified @actor and
* passing the @user_data pointer each time.
*
* Return value: returns %TRUE if all children were iterated, else
* %FALSE if a callback broke out of iteration early.
*/
gboolean
_clutter_actor_foreach_child (ClutterActor *self,
ClutterForeachCallback callback,
gpointer user_data)
{
ClutterActorPrivate *priv = self->priv;
ClutterActor *iter;
gboolean cont;
for (cont = TRUE, iter = priv->first_child;
cont && iter != NULL;
iter = iter->priv->next_sibling)
{
cont = callback (iter, user_data);
}
return cont;
}
/* For debugging purposes this gives us a simple way to print out
* the scenegraph e.g in gdb using:
* [|
* _clutter_actor_traverse (stage,
* 0,
* _clutter_debug_print_actor_cb,
* NULL,
* NULL);
* |]
*/
ClutterActorTraverseVisitFlags
_clutter_debug_print_actor_cb (ClutterActor *actor,
int depth,
void *user_data)
{
g_print ("%*s%s:%p\n",
depth * 2, "",
_clutter_actor_get_debug_name (actor),
actor);
return CLUTTER_ACTOR_TRAVERSE_VISIT_CONTINUE;
}
static void
_clutter_actor_traverse_breadth (ClutterActor *actor,
ClutterTraverseCallback callback,
gpointer user_data)
{
GQueue *queue = g_queue_new ();
ClutterActor dummy;
int current_depth = 0;
g_queue_push_tail (queue, actor);
g_queue_push_tail (queue, &dummy); /* use to delimit depth changes */
while ((actor = g_queue_pop_head (queue)))
{
ClutterActorTraverseVisitFlags flags;
if (actor == &dummy)
{
current_depth++;
g_queue_push_tail (queue, &dummy);
continue;
}
flags = callback (actor, current_depth, user_data);
if (flags & CLUTTER_ACTOR_TRAVERSE_VISIT_BREAK)
break;
else if (!(flags & CLUTTER_ACTOR_TRAVERSE_VISIT_SKIP_CHILDREN))
{
ClutterActor *iter;
for (iter = actor->priv->first_child;
iter != NULL;
iter = iter->priv->next_sibling)
{
g_queue_push_tail (queue, iter);
}
}
}
g_queue_free (queue);
}
static ClutterActorTraverseVisitFlags
_clutter_actor_traverse_depth (ClutterActor *actor,
ClutterTraverseCallback before_children_callback,
ClutterTraverseCallback after_children_callback,
int current_depth,
gpointer user_data)
{
ClutterActorTraverseVisitFlags flags;
flags = before_children_callback (actor, current_depth, user_data);
if (flags & CLUTTER_ACTOR_TRAVERSE_VISIT_BREAK)
return CLUTTER_ACTOR_TRAVERSE_VISIT_BREAK;
if (!(flags & CLUTTER_ACTOR_TRAVERSE_VISIT_SKIP_CHILDREN))
{
ClutterActor *iter;
for (iter = actor->priv->first_child;
iter != NULL;
iter = iter->priv->next_sibling)
{
flags = _clutter_actor_traverse_depth (iter,
before_children_callback,
after_children_callback,
current_depth + 1,
user_data);
if (flags & CLUTTER_ACTOR_TRAVERSE_VISIT_BREAK)
return CLUTTER_ACTOR_TRAVERSE_VISIT_BREAK;
}
}
if (after_children_callback)
return after_children_callback (actor, current_depth, user_data);
else
return CLUTTER_ACTOR_TRAVERSE_VISIT_CONTINUE;
}
/* _clutter_actor_traverse:
* @actor: The actor to start traversing the graph from
* @flags: These flags may affect how the traversal is done
* @before_children_callback: A function to call before visiting the
* children of the current actor.
* @after_children_callback: A function to call after visiting the
* children of the current actor. (Ignored if
* %CLUTTER_ACTOR_TRAVERSE_BREADTH_FIRST is passed to @flags.)
* @user_data: The private data to pass to the callbacks
*
* Traverses the scenegraph starting at the specified @actor and
* descending through all its children and its children's children.
* For each actor traversed @before_children_callback and
* @after_children_callback are called with the specified
* @user_data, before and after visiting that actor's children.
*
* The callbacks can return flags that affect the ongoing traversal
* such as by skipping over an actors children or bailing out of
* any further traversing.
*/
void
_clutter_actor_traverse (ClutterActor *actor,
ClutterActorTraverseFlags flags,
ClutterTraverseCallback before_children_callback,
ClutterTraverseCallback after_children_callback,
gpointer user_data)
{
if (flags & CLUTTER_ACTOR_TRAVERSE_BREADTH_FIRST)
_clutter_actor_traverse_breadth (actor,
before_children_callback,
user_data);
else /* DEPTH_FIRST */
_clutter_actor_traverse_depth (actor,
before_children_callback,
after_children_callback,
0, /* start depth */
user_data);
}
static void
on_layout_manager_changed (ClutterLayoutManager *manager,
ClutterActor *self)
{
clutter_actor_queue_relayout (self);
}
2012-01-03 14:31:06 +00:00
/**
* clutter_actor_set_layout_manager:
* @self: a #ClutterActor
* @manager: (allow-none): a #ClutterLayoutManager, or %NULL to unset it
*
* Sets the #ClutterLayoutManager delegate object that will be used to
* lay out the children of @self.
*
* The #ClutterActor will take a reference on the passed @manager which
* will be released either when the layout manager is removed, or when
* the actor is destroyed.
*
* Since: 1.10
*/
void
clutter_actor_set_layout_manager (ClutterActor *self,
ClutterLayoutManager *manager)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (manager == NULL || CLUTTER_IS_LAYOUT_MANAGER (manager));
priv = self->priv;
if (priv->layout_manager != NULL)
{
g_signal_handlers_disconnect_by_func (priv->layout_manager,
G_CALLBACK (on_layout_manager_changed),
self);
clutter_layout_manager_set_container (priv->layout_manager, NULL);
g_object_unref (priv->layout_manager);
}
priv->layout_manager = manager;
if (priv->layout_manager != NULL)
{
g_object_ref_sink (priv->layout_manager);
clutter_layout_manager_set_container (priv->layout_manager,
CLUTTER_CONTAINER (self));
g_signal_connect (priv->layout_manager, "layout-changed",
G_CALLBACK (on_layout_manager_changed),
self);
}
clutter_actor_queue_relayout (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_LAYOUT_MANAGER]);
}
/**
* clutter_actor_get_layout_manager:
* @self: a #ClutterActor
*
* Retrieves the #ClutterLayoutManager used by @self.
*
* Return value: (transfer none): a pointer to the #ClutterLayoutManager,
* or %NULL
*
* Since: 1.10
*/
ClutterLayoutManager *
clutter_actor_get_layout_manager (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
return self->priv->layout_manager;
}
2011-11-20 22:00:01 +00:00
static const ClutterLayoutInfo default_layout_info = {
0.f, /* fixed-x */
0.f, /* fixed-y */
{ 0, 0, 0, 0 }, /* margin */
CLUTTER_ACTOR_ALIGN_FILL, /* x-align */
CLUTTER_ACTOR_ALIGN_FILL, /* y-align */
0.f, 0.f, /* min_width, natural_width */
0.f, 0.f, /* natual_width, natural_height */
};
static void
layout_info_free (gpointer data)
{
if (G_LIKELY (data != NULL))
g_slice_free (ClutterLayoutInfo, data);
}
/*< private >
* _clutter_actor_get_layout_info:
* @self: a #ClutterActor
*
* Retrieves a pointer to the ClutterLayoutInfo structure.
*
* If the actor does not have a ClutterLayoutInfo associated to it, one
* will be created and initialized to the default values.
*
* This function should be used for setters.
*
* For getters, you should use _clutter_actor_get_layout_info_or_defaults()
* instead.
*
* Return value: (transfer none): a pointer to the ClutterLayoutInfo structure
*/
ClutterLayoutInfo *
_clutter_actor_get_layout_info (ClutterActor *self)
{
ClutterLayoutInfo *retval;
retval = g_object_get_qdata (G_OBJECT (self), quark_actor_layout_info);
if (retval == NULL)
{
retval = g_slice_new (ClutterLayoutInfo);
*retval = default_layout_info;
g_object_set_qdata_full (G_OBJECT (self), quark_actor_layout_info,
retval,
layout_info_free);
}
return retval;
}
/*< private >
* _clutter_actor_get_layout_info_or_defaults:
* @self: a #ClutterActor
*
* Retrieves the ClutterLayoutInfo structure associated to an actor.
*
* If the actor does not have a ClutterLayoutInfo structure associated to it,
* then the default structure will be returned.
*
* This function should only be used for getters.
*
* Return value: a const pointer to the ClutterLayoutInfo structure
*/
const ClutterLayoutInfo *
_clutter_actor_get_layout_info_or_defaults (ClutterActor *self)
{
const ClutterLayoutInfo *info;
info = g_object_get_qdata (G_OBJECT (self), quark_actor_layout_info);
if (info == NULL)
return &default_layout_info;
return info;
}
/**
* clutter_actor_set_x_align:
* @self: a #ClutterActor
* @x_align: the horizontal alignment policy
*
* Sets the horizontal alignment policy of a #ClutterActor, in case the
* actor received extra horizontal space.
*
* See also the #ClutterActor:x-align property.
*
* Since: 1.10
*/
void
clutter_actor_set_x_align (ClutterActor *self,
ClutterActorAlign x_align)
{
ClutterLayoutInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
info = _clutter_actor_get_layout_info (self);
if (info->x_align != x_align)
{
info->x_align = x_align;
clutter_actor_queue_relayout (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_X_ALIGN]);
}
}
/**
* clutter_actor_get_x_align:
* @self: a #ClutterActor
*
* Retrieves the horizontal alignment policy set using
* clutter_actor_set_x_align().
*
* Return value: the horizontal alignment policy.
*
* Since: 1.10
*/
ClutterActorAlign
clutter_actor_get_x_align (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), CLUTTER_ACTOR_ALIGN_FILL);
return _clutter_actor_get_layout_info_or_defaults (self)->x_align;
}
/**
* clutter_actor_set_y_align:
* @self: a #ClutterActor
* @y_align: the vertical alignment policy
*
* Sets the vertical alignment policy of a #ClutterActor, in case the
* actor received extra vertical space.
*
* See also the #ClutterActor:y-align property.
*
* Since: 1.10
*/
void
clutter_actor_set_y_align (ClutterActor *self,
ClutterActorAlign y_align)
{
ClutterLayoutInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
info = _clutter_actor_get_layout_info (self);
if (info->y_align != y_align)
{
info->y_align = y_align;
clutter_actor_queue_relayout (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_Y_ALIGN]);
}
}
/**
* clutter_actor_get_y_align:
* @self: a #ClutterActor
*
* Retrieves the vertical alignment policy set using
* clutter_actor_set_y_align().
*
* Return value: the vertical alignment policy.
*
* Since: 1.10
*/
ClutterActorAlign
clutter_actor_get_y_align (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), CLUTTER_ACTOR_ALIGN_FILL);
return _clutter_actor_get_layout_info_or_defaults (self)->y_align;
}
/**
* clutter_margin_new:
*
* Creates a new #ClutterMargin.
*
* Return value: (transfer full): a newly allocated #ClutterMargin. Use
* clutter_margin_free() to free the resources associated with it when
* done.
*
* Since: 1.10
*/
ClutterMargin *
clutter_margin_new (void)
{
return g_slice_new0 (ClutterMargin);
}
/**
* clutter_margin_copy:
* @margin_: a #ClutterMargin
*
* Creates a new #ClutterMargin and copies the contents of @margin_ into
* the newly created structure.
*
* Return value: (transfer full): a copy of the #ClutterMargin.
*
* Since: 1.10
*/
ClutterMargin *
clutter_margin_copy (const ClutterMargin *margin_)
{
if (G_LIKELY (margin_ != NULL))
return g_slice_dup (ClutterMargin, margin_);
return NULL;
}
/**
* clutter_margin_free:
* @margin_: a #ClutterMargin
*
* Frees the resources allocated by clutter_margin_new() and
* clutter_margin_copy().
*
* Since: 1.10
*/
void
clutter_margin_free (ClutterMargin *margin_)
{
if (G_LIKELY (margin_ != NULL))
g_slice_free (ClutterMargin, margin_);
}
G_DEFINE_BOXED_TYPE (ClutterMargin, clutter_margin,
clutter_margin_copy,
clutter_margin_free)
/**
* clutter_actor_set_margin:
* @self: a #ClutterActor
* @margin: a #ClutterMargin
*
* Sets all the components of the margin of a #ClutterActor.
*
* Since: 1.10
*/
void
clutter_actor_set_margin (ClutterActor *self,
const ClutterMargin *margin)
{
ClutterLayoutInfo *info;
gboolean changed;
GObject *obj;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (margin != NULL);
obj = G_OBJECT (self);
changed = FALSE;
g_object_freeze_notify (obj);
info = _clutter_actor_get_layout_info (self);
if (info->margin.top != margin->top)
{
info->margin.top = margin->top;
g_object_notify_by_pspec (obj, obj_props[PROP_MARGIN_TOP]);
changed = TRUE;
}
if (info->margin.right != margin->right)
{
info->margin.right = margin->right;
g_object_notify_by_pspec (obj, obj_props[PROP_MARGIN_RIGHT]);
changed = TRUE;
}
if (info->margin.bottom != margin->bottom)
{
info->margin.bottom = margin->bottom;
g_object_notify_by_pspec (obj, obj_props[PROP_MARGIN_BOTTOM]);
changed = TRUE;
}
if (info->margin.left != margin->left)
{
info->margin.left = margin->left;
g_object_notify_by_pspec (obj, obj_props[PROP_MARGIN_LEFT]);
changed = TRUE;
}
if (changed)
clutter_actor_queue_relayout (self);
g_object_thaw_notify (obj);
}
/**
* clutter_actor_get_margin:
* @self: a #ClutterActor
* @margin: (out caller-allocates): return location for a #ClutterMargin
*
* Retrieves all the components of the margin of a #ClutterActor.
*
* Since: 1.10
*/
void
clutter_actor_get_margin (ClutterActor *self,
ClutterMargin *margin)
{
const ClutterLayoutInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (margin != NULL);
info = _clutter_actor_get_layout_info_or_defaults (self);
*margin = info->margin;
}
/**
* clutter_actor_set_margin_top:
* @self: a #ClutterActor
* @margin: the top margin
*
* Sets the margin from the top of a #ClutterActor.
*
* Since: 1.10
*/
void
clutter_actor_set_margin_top (ClutterActor *self,
gfloat margin)
{
ClutterLayoutInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (margin >= 0.f);
info = _clutter_actor_get_layout_info (self);
if (info->margin.top == margin)
return;
info->margin.top = margin;
clutter_actor_queue_relayout (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MARGIN_TOP]);
}
/**
* clutter_actor_get_margin_top:
* @self: a #ClutterActor
*
* Retrieves the top margin of a #ClutterActor.
*
* Return value: the top margin
*
* Since: 1.10
*/
gfloat
clutter_actor_get_margin_top (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0.f);
return _clutter_actor_get_layout_info_or_defaults (self)->margin.top;
}
/**
* clutter_actor_set_margin_bottom:
* @self: a #ClutterActor
* @margin: the bottom margin
*
* Sets the margin from the bottom of a #ClutterActor.
*
* Since: 1.10
*/
void
clutter_actor_set_margin_bottom (ClutterActor *self,
gfloat margin)
{
ClutterLayoutInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (margin >= 0.f);
info = _clutter_actor_get_layout_info (self);
if (info->margin.bottom == margin)
return;
info->margin.bottom = margin;
clutter_actor_queue_relayout (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MARGIN_BOTTOM]);
}
/**
* clutter_actor_get_margin_bottom:
* @self: a #ClutterActor
*
* Retrieves the bottom margin of a #ClutterActor.
*
* Return value: the bottom margin
*
* Since: 1.10
*/
gfloat
clutter_actor_get_margin_bottom (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0.f);
return _clutter_actor_get_layout_info_or_defaults (self)->margin.bottom;
}
/**
* clutter_actor_set_margin_left:
* @self: a #ClutterActor
* @margin: the left margin
*
* Sets the margin from the left of a #ClutterActor.
*
* Since: 1.10
*/
void
clutter_actor_set_margin_left (ClutterActor *self,
gfloat margin)
{
ClutterLayoutInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (margin >= 0.f);
info = _clutter_actor_get_layout_info (self);
if (info->margin.left == margin)
return;
info->margin.left = margin;
clutter_actor_queue_relayout (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MARGIN_LEFT]);
}
/**
* clutter_actor_get_margin_left:
* @self: a #ClutterActor
*
* Retrieves the left margin of a #ClutterActor.
*
* Return value: the left margin
*
* Since: 1.10
*/
gfloat
clutter_actor_get_margin_left (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0.f);
return _clutter_actor_get_layout_info_or_defaults (self)->margin.left;
}
/**
* clutter_actor_set_margin_right:
* @self: a #ClutterActor
* @margin: the right margin
*
* Sets the margin from the right of a #ClutterActor.
*
* Since: 1.10
*/
void
clutter_actor_set_margin_right (ClutterActor *self,
gfloat margin)
{
ClutterLayoutInfo *info;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (margin >= 0.f);
info = _clutter_actor_get_layout_info (self);
if (info->margin.right == margin)
return;
info->margin.right = margin;
clutter_actor_queue_relayout (self);
g_object_notify_by_pspec (G_OBJECT (self), obj_props[PROP_MARGIN_RIGHT]);
}
/**
* clutter_actor_get_margin_right:
* @self: a #ClutterActor
*
* Retrieves the right margin of a #ClutterActor.
*
* Return value: the right margin
*
* Since: 1.10
*/
gfloat
clutter_actor_get_margin_right (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), 0.f);
return _clutter_actor_get_layout_info_or_defaults (self)->margin.right;
}
/**
* clutter_actor_set_background_color:
* @self: a #ClutterActor
* @color: (allow-none): a #ClutterColor, or %NULL to unset a previously
* set color
*
* Sets the background color of a #ClutterActor.
*
* The background color will be used to cover the whole allocation of the
* actor. The default background color of an actor is transparent.
*
* To check whether an actor has a background color, you can use the
* #ClutterActor:background-color-set actor property.
*
* Since: 1.10
*/
void
clutter_actor_set_background_color (ClutterActor *self,
const ClutterColor *color)
{
ClutterActorPrivate *priv;
g_return_if_fail (CLUTTER_IS_ACTOR (self));
priv = self->priv;
if (color == NULL)
{
priv->bg_color_set = FALSE;
g_object_notify_by_pspec (G_OBJECT (self),
obj_props[PROP_BACKGROUND_COLOR_SET]);
return;
}
if (priv->bg_color_set && clutter_color_equal (color, &priv->bg_color))
return;
priv->bg_color = *color;
priv->bg_color_set = TRUE;
clutter_actor_queue_redraw (self);
g_object_notify_by_pspec (G_OBJECT (self),
obj_props[PROP_BACKGROUND_COLOR_SET]);
g_object_notify_by_pspec (G_OBJECT (self),
obj_props[PROP_BACKGROUND_COLOR]);
}
/**
* clutter_actor_get_background_color:
* @self: a #ClutterActor
* @color: (out caller-allocates): return location for a #ClutterColor
*
* Retrieves the color set using clutter_actor_set_background_color().
*
* Since: 1.10
*/
void
clutter_actor_get_background_color (ClutterActor *self,
ClutterColor *color)
{
g_return_if_fail (CLUTTER_IS_ACTOR (self));
g_return_if_fail (color != NULL);
*color = self->priv->bg_color;
}
/**
* clutter_actor_get_previous_sibling:
* @self: a #ClutterActor
*
* Retrieves the sibling of @self that comes before it in the list
* of children of @self's parent.
*
* The returned pointer is only valid until the scene graph changes; it
* is not safe to modify the list of children of @self while iterating
* it.
*
* Return value: (transfer none): a pointer to a #ClutterActor, or %NULL
*
* Since: 1.10
*/
ClutterActor *
clutter_actor_get_previous_sibling (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
return self->priv->prev_sibling;
}
/**
* clutter_actor_get_next_sibling:
* @self: a #ClutterActor
*
* Retrieves the sibling of @self that comes after it in the list
* of children of @self's parent.
*
* The returned pointer is only valid until the scene graph changes; it
* is not safe to modify the list of children of @self while iterating
* it.
*
* Return value: (transfer none): a pointer to a #ClutterActor, or %NULL
*
* Since: 1.10
*/
ClutterActor *
clutter_actor_get_next_sibling (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
return self->priv->next_sibling;
}
/**
* clutter_actor_get_first_child:
* @self: a #ClutterActor
*
* Retrieves the first child of @self.
*
* The returned pointer is only valid until the scene graph changes; it
* is not safe to modify the list of children of @self while iterating
* it.
*
* Return value: (transfer none): a pointer to a #ClutterActor, or %NULL
*
* Since: 1.10
*/
ClutterActor *
clutter_actor_get_first_child (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
return self->priv->first_child;
}
/**
* clutter_actor_get_last_child:
* @self: a #ClutterActor
*
* Retrieves the last child of @self.
*
* The returned pointer is only valid until the scene graph changes; it
* is not safe to modify the list of children of @self while iterating
* it.
*
* Return value: (transfer none): a pointer to a #ClutterActor, or %NULL
*
* Since: 1.10
*/
ClutterActor *
clutter_actor_get_last_child (ClutterActor *self)
{
g_return_val_if_fail (CLUTTER_IS_ACTOR (self), NULL);
return self->priv->last_child;
}
/* easy way to have properly named fields instead of the dummy ones
* we use in the public structure
*/
typedef struct _RealActorIter
{
ClutterActor *root; /* dummy1 */
ClutterActor *current; /* dummy2 */
gpointer padding_1; /* dummy3 */
gint age; /* dummy4 */
gpointer padding_2; /* dummy5 */
} RealActorIter;
/**
* clutter_actor_iter_init:
* @iter: a #ClutterActorIter
* @root: a #ClutterActor
*
* Initializes a #ClutterActorIter, which can then be used to iterate
* efficiently over a section of the scene graph, and associates it
* with @root.
*
* Modifying the scene graph section that contains @root will invalidate
* the iterator.
*
* |[
* ClutterActorIter iter;
* ClutterActor *child;
*
* clutter_actor_iter_init (&iter, container);
* while (clutter_actor_iter_next (&iter, &child))
* {
* /&ast; do something with child &ast;/
* }
* ]|
*
* Since: 1.10
*/
void
clutter_actor_iter_init (ClutterActorIter *iter,
ClutterActor *root)
{
RealActorIter *ri = (RealActorIter *) iter;
g_return_if_fail (iter != NULL);
g_return_if_fail (CLUTTER_IS_ACTOR (root));
ri->root = root;
ri->current = NULL;
ri->age = root->priv->age;
}
/**
* clutter_actor_iter_next:
* @iter: a #ClutterActorIter
* @child: (out): return location for a #ClutterActor
*
* Advances the @iter and retrieves the next child of the root #ClutterActor
* that was used to initialize the #ClutterActorIterator.
*
* If the iterator can advance, this function returns %TRUE and sets the
* @child argument.
*
* If the iterator cannot advance, this function returns %FALSE, and
* the contents of @child are undefined.
*
* Return value: %TRUE if the iterator could advance, and %FALSE otherwise.
*
* Since: 1.10
*/
gboolean
clutter_actor_iter_next (ClutterActorIter *iter,
ClutterActor **child)
{
RealActorIter *ri = (RealActorIter *) iter;
g_return_val_if_fail (iter != NULL, FALSE);
g_return_val_if_fail (ri->root != NULL, FALSE);
#ifndef G_DISABLE_ASSERT
g_return_val_if_fail (ri->age == ri->root->priv->age, FALSE);
#endif
if (ri->current == NULL)
ri->current = ri->root->priv->first_child;
else
ri->current = ri->current->priv->next_sibling;
if (child != NULL)
*child = ri->current;
return ri->current != NULL;
}
/**
* clutter_actor_iter_next:
* @iter: a #ClutterActorIter
* @child: (out): return location for a #ClutterActor
*
* Advances the @iter and retrieves the previous child of the root
* #ClutterActor that was used to initialize the #ClutterActorIterator.
*
* If the iterator can advance, this function returns %TRUE and sets the
* @child argument.
*
* If the iterator cannot advance, this function returns %FALSE, and
* the contents of @child are undefined.
*
* Return value: %TRUE if the iterator could advance, and %FALSE otherwise.
*
* Since: 1.10
*/
gboolean
clutter_actor_iter_prev (ClutterActorIter *iter,
ClutterActor **child)
{
RealActorIter *ri = (RealActorIter *) iter;
g_return_val_if_fail (iter != NULL, FALSE);
g_return_val_if_fail (ri->root != NULL, FALSE);
#ifndef G_DISABLE_ASSERT
g_return_val_if_fail (ri->age == ri->root->priv->age, FALSE);
#endif
if (ri->current == NULL)
ri->current = ri->root->priv->last_child;
else
ri->current = ri->current->priv->prev_sibling;
if (child != NULL)
*child = ri->current;
return ri->current != NULL;
}
/**
* clutter_actor_iter_remove:
* @iter: a #ClutterActorIter
*
* Safely removes the #ClutterActor currently pointer to by the iterator
* from its parent.
*
* This function can only be called after clutter_actor_iter_next() or
* clutter_actor_iter_prev() returned %TRUE, and cannot be called more
* than once for the same actor.
*
* This function will call clutter_actor_remove_child() internally.
*
* Since: 1.10
*/
void
clutter_actor_iter_remove (ClutterActorIter *iter)
{
RealActorIter *ri = (RealActorIter *) iter;
ClutterActor *cur;
g_return_if_fail (iter != NULL);
g_return_if_fail (ri->root != NULL);
#ifndef G_DISABLE_ASSERT
g_return_if_fail (ri->age == ri->root->priv->age);
#endif
g_return_if_fail (ri->current != NULL);
cur = ri->current;
if (cur != NULL)
{
ri->current = cur->priv->prev_sibling;
clutter_actor_remove_child_internal (ri->root, cur,
REMOVE_CHILD_DEFAULT_FLAGS);
ri->age += 1;
}
}
/**
* clutter_actor_iter_destroy:
* @iter: a #ClutterActorIter
*
* Safely destroys the #ClutterActor currently pointer to by the iterator
* from its parent.
*
* This function can only be called after clutter_actor_iter_next() or
* clutter_actor_iter_prev() returned %TRUE, and cannot be called more
* than once for the same actor.
*
* This function will call clutter_actor_destroy() internally.
*
* Since: 1.10
*/
void
clutter_actor_iter_destroy (ClutterActorIter *iter)
{
RealActorIter *ri = (RealActorIter *) iter;
ClutterActor *cur;
g_return_if_fail (iter != NULL);
g_return_if_fail (ri->root != NULL);
#ifndef G_DISABLE_ASSERT
g_return_if_fail (ri->age == ri->root->priv->age);
#endif
g_return_if_fail (ri->current != NULL);
cur = ri->current;
if (cur != NULL)
{
ri->current = cur->priv->prev_sibling;
clutter_actor_destroy (cur);
ri->age += 1;
}
}