mutter/clutter/clutter-effect.c

401 lines
14 KiB
C
Raw Normal View History

/*
* Clutter.
*
* An OpenGL based 'interactive canvas' library.
*
* Copyright (C) 2010 Intel Corporation.
*
* 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/>.
*
* Author:
* Emmanuele Bassi <ebassi@linux.intel.com>
*/
/**
* SECTION:clutter-effect
* @short_description: Base class for actor effects
*
* The #ClutterEffect class provides a default type and API for creating
* effects for generic actors.
*
* Effects are a #ClutterActorMeta sub-class that modify the way an actor
* is painted in a way that is not part of the actor's implementation.
*
* Effects should be the preferred way to affect the paint sequence of an
* actor without sub-classing the actor itself and overriding the
* #ClutterActorClass.paint()_ virtual function.
*
* <refsect2 id="ClutterEffect-implementation">
* <title>Implementing a ClutterEffect</title>
* <para>
* Creating a sub-class of #ClutterEffect requires overriding the
* paint method. The implementation of the function should look
* something like this:
* </para>
* <programlisting>
* void effect_paint (ClutterEffect *effect, ClutterEffectPaintFlags flags)
* {
* /&ast; Set up initialisation of the paint such as binding a
* CoglOffscreen or other operations &ast;/
*
* /&ast; Chain to the next item in the paint sequence. This will either call
* paint on the next effect or just paint the actor if this is
* the last effect. &ast;/
* ClutterActor *actor =
* clutter_actor_meta_get_actor (CLUTTER_ACTOR_META (effect));
* clutter_actor_continue_paint (actor);
*
* /&ast; perform any cleanup of state, such as popping the
* CoglOffscreen &ast;/
* }
* </programlisting>
* <para>
* The effect can optionally avoid calling
* clutter_actor_continue_paint() to skip any further stages of
* the paint sequence. This is useful for example if the effect
* contains a cached image of the actor. In that case it can
* optimise painting by avoiding the actor paint and instead
2011-09-12 08:12:14 -04:00
* painting the cached image. The %CLUTTER_EFFECT_PAINT_ACTOR_DIRTY
* flag is useful in this case. Clutter will set this flag when a
* redraw has been queued on the actor since it was last
* painted. The effect can use this information to decide if the
* cached image is still valid.
* </para>
* <para>
* The paint virtual was added in Clutter 1.8. Prior to that there
* were two separate functions as follows.
* </para>
* <itemizedlist>
* <listitem><simpara><function>pre_paint()</function>, which is called
* before painting the #ClutterActor.</simpara></listitem>
* <listitem><simpara><function>post_paint()</function>, which is called
* after painting the #ClutterActor.</simpara></listitem>
* </itemizedlist>
* <para>The <function>pre_paint()</function> function was used to set
* up the #ClutterEffect right before the #ClutterActor's paint
* sequence. This function can fail, and return %FALSE; in that case, no
* <function>post_paint()</function> invocation will follow.</para>
* <para>The <function>post_paint()</function> function was called after the
* #ClutterActor's paint sequence.</para>
* <para>
* With these two functions it is not possible to skip the rest of
* the paint sequence. The default implementation of the paint
* virtual calls #ClutterEffectClass.pre_paint(), clutter_actor_continue_paint()
* and then #ClutterEffectClass.post_paint() so that existing actors that aren't
* using the #ClutterEffectClass.paint() virtual will continue to work. New
* effects using the #ClutterEffectClass.paint() virtual do not need to implement
* pre or post paint.
* </para>
* <example id="ClutterEffect-example">
* <title>A simple ClutterEffect implementation</title>
* <para>The example below creates two rectangles: one will be
* painted "behind" the actor, while another will be painted "on
* top" of the actor. The #ClutterActorMetaClass.set_actor()
* implementation will create the two materials used for the two
* different rectangles; the #ClutterEffectClass.paint() implementation
* will paint the first material using cogl_rectangle(), before
* continuing and then it will paint paint the second material
* after.</para>
* <programlisting>
* typedef struct {
* ClutterEffect parent_instance;
*
* CoglHandle rect_1;
* CoglHandle rect_2;
* } MyEffect;
*
* typedef struct _ClutterEffectClass MyEffectClass;
*
* G_DEFINE_TYPE (MyEffect, my_effect, CLUTTER_TYPE_EFFECT);
*
* static void
* my_effect_set_actor (ClutterActorMeta *meta,
* ClutterActor *actor)
* {
* MyEffect *self = MY_EFFECT (meta);
*
* /&ast; Clear the previous state &ast;/
* if (self-&gt;rect_1)
* {
* cogl_handle_unref (self-&gt;rect_1);
* self-&gt;rect_1 = NULL;
* }
*
* if (self-&gt;rect_2)
* {
* cogl_handle_unref (self-&gt;rect_2);
* self-&gt;rect_2 = NULL;
* }
*
* /&ast; Maintain a pointer to the actor &ast;
* self-&gt;actor = actor;
*
* /&ast; If we've been detached by the actor then we should
* &ast; just bail out here
* &ast;/
* if (self-&gt;actor == NULL)
* return;
*
* /&ast; Create a red material &ast;/
* self-&gt;rect_1 = cogl_material_new ();
* cogl_material_set_color4f (self-&gt;rect_1, 1.0, 0.0, 0.0, 1.0);
*
* /&ast; Create a green material &ast;/
* self-&gt;rect_2 = cogl_material_new ();
* cogl_material_set_color4f (self-&gt;rect_2, 0.0, 1.0, 0.0, 1.0);
* }
*
* static gboolean
* my_effect_paint (ClutterEffect *effect)
* {
* MyEffect *self = MY_EFFECT (effect);
* gfloat width, height;
*
* clutter_actor_get_size (self-&gt;actor, &amp;width, &amp;height);
*
* /&ast; Paint the first rectangle in the upper left quadrant &ast;/
* cogl_set_source (self-&gt;rect_1);
* cogl_rectangle (0, 0, width / 2, height / 2);
*
* /&ast; Continue to the rest of the paint sequence &ast;/
* clutter_actor_continue_paint (self-&gt;actor);
*
* /&ast; Paint the second rectangle in the lower right quadrant &ast;/
* cogl_set_source (self-&gt;rect_2);
* cogl_rectangle (width / 2, height / 2, width, height);
* }
*
* static void
* my_effect_class_init (MyEffectClass *klass)
* {
* ClutterActorMetaClas *meta_class = CLUTTER_ACTOR_META_CLASS (klass);
*
* meta_class-&gt;set_actor = my_effect_set_actor;
*
* klass-&gt;paint = my_effect_paint;
* }
* </programlisting>
* </example>
* </refsect2>
*
* #ClutterEffect is available since Clutter 1.4
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "clutter-effect.h"
#include "clutter-actor-meta-private.h"
#include "clutter-debug.h"
#include "clutter-effect-private.h"
#include "clutter-enum-types.h"
#include "clutter-marshal.h"
#include "clutter-private.h"
#include "clutter-actor-private.h"
G_DEFINE_ABSTRACT_TYPE (ClutterEffect,
clutter_effect,
CLUTTER_TYPE_ACTOR_META);
static gboolean
clutter_effect_real_pre_paint (ClutterEffect *effect)
{
return TRUE;
}
static void
clutter_effect_real_post_paint (ClutterEffect *effect)
{
}
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 13:04:19 -04:00
static gboolean
clutter_effect_real_get_paint_volume (ClutterEffect *effect,
ClutterPaintVolume *volume)
{
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 13:04:19 -04:00
return TRUE;
}
static void
clutter_effect_real_paint (ClutterEffect *effect,
ClutterEffectPaintFlags flags)
{
ClutterActorMeta *actor_meta = CLUTTER_ACTOR_META (effect);
ClutterActor *actor;
gboolean pre_paint_succeeded;
/* The default implementation provides a compatibility wrapper for
effects that haven't migrated to use the 'paint' virtual yet. This
just calls the old pre and post virtuals before chaining on */
pre_paint_succeeded = _clutter_effect_pre_paint (effect);
actor = clutter_actor_meta_get_actor (actor_meta);
clutter_actor_continue_paint (actor);
if (pre_paint_succeeded)
_clutter_effect_post_paint (effect);
}
static void
clutter_effect_real_pick (ClutterEffect *effect,
ClutterEffectPaintFlags flags)
{
ClutterActorMeta *actor_meta = CLUTTER_ACTOR_META (effect);
ClutterActor *actor;
actor = clutter_actor_meta_get_actor (actor_meta);
clutter_actor_continue_paint (actor);
}
static void
clutter_effect_notify (GObject *gobject,
GParamSpec *pspec)
{
if (strcmp (pspec->name, "enabled") == 0)
{
ClutterActorMeta *meta = CLUTTER_ACTOR_META (gobject);
ClutterActor *actor = clutter_actor_meta_get_actor (meta);
if (actor != NULL)
clutter_actor_queue_redraw (actor);
}
if (G_OBJECT_CLASS (clutter_effect_parent_class)->notify != NULL)
G_OBJECT_CLASS (clutter_effect_parent_class)->notify (gobject, pspec);
}
static void
clutter_effect_class_init (ClutterEffectClass *klass)
{
GObjectClass *gobject_class = G_OBJECT_CLASS (klass);
gobject_class->notify = clutter_effect_notify;
klass->pre_paint = clutter_effect_real_pre_paint;
klass->post_paint = clutter_effect_real_post_paint;
klass->get_paint_volume = clutter_effect_real_get_paint_volume;
klass->paint = clutter_effect_real_paint;
klass->pick = clutter_effect_real_pick;
}
static void
clutter_effect_init (ClutterEffect *self)
{
}
gboolean
_clutter_effect_pre_paint (ClutterEffect *effect)
{
g_return_val_if_fail (CLUTTER_IS_EFFECT (effect), FALSE);
return CLUTTER_EFFECT_GET_CLASS (effect)->pre_paint (effect);
}
void
_clutter_effect_post_paint (ClutterEffect *effect)
{
g_return_if_fail (CLUTTER_IS_EFFECT (effect));
CLUTTER_EFFECT_GET_CLASS (effect)->post_paint (effect);
}
void
_clutter_effect_paint (ClutterEffect *effect,
ClutterEffectPaintFlags flags)
{
g_return_if_fail (CLUTTER_IS_EFFECT (effect));
CLUTTER_EFFECT_GET_CLASS (effect)->paint (effect, flags);
}
void
_clutter_effect_pick (ClutterEffect *effect,
ClutterEffectPaintFlags flags)
{
g_return_if_fail (CLUTTER_IS_EFFECT (effect));
CLUTTER_EFFECT_GET_CLASS (effect)->pick (effect, flags);
}
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 13:04:19 -04:00
gboolean
_clutter_effect_get_paint_volume (ClutterEffect *effect,
ClutterPaintVolume *volume)
{
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 13:04:19 -04:00
g_return_val_if_fail (CLUTTER_IS_EFFECT (effect), FALSE);
g_return_val_if_fail (volume != NULL, 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 13:04:19 -04:00
return CLUTTER_EFFECT_GET_CLASS (effect)->get_paint_volume (effect, volume);
}
/**
* clutter_effect_queue_repaint:
* @effect: A #ClutterEffect which needs redrawing
*
* Queues a repaint of the effect. The effect can detect when the paint
* method is called as a result of this function because it will not
2011-09-12 08:12:14 -04:00
* have the %CLUTTER_EFFECT_PAINT_ACTOR_DIRTY flag set. In that case the
* effect is free to assume that the actor has not changed its
* appearance since the last time it was painted so it doesn't need to
* call clutter_actor_continue_paint() if it can draw a cached
* image. This is mostly intended for effects that are using a
* %CoglOffscreen to redirect the actor (such as
* %ClutterOffscreenEffect). In that case the effect can save a bit of
* rendering time by painting the cached texture without causing the
* entire actor to be painted.
*
* This function can be used by effects that have their own animatable
* parameters. For example, an effect which adds a varying degree of a
* red tint to an actor by redirecting it through a CoglOffscreen
* might have a property to specify the level of tint. When this value
* changes, the underlying actor doesn't need to be redrawn so the
* effect can call clutter_effect_queue_repaint() to make sure the
* effect is repainted.
*
* Note however that modifying the position of the parent of an actor
* may change the appearance of the actor because its transformation
* matrix would change. In this case a redraw wouldn't be queued on
2011-09-12 08:12:14 -04:00
* the actor itself so the %CLUTTER_EFFECT_PAINT_ACTOR_DIRTY would still
* not be set. The effect can detect this case by keeping track of the
* last modelview matrix that was used to render the actor and
* veryifying that it remains the same in the next paint.
*
* Any other effects that are layered on top of the passed in effect
2011-09-12 08:12:14 -04:00
* will still be passed the %CLUTTER_EFFECT_PAINT_ACTOR_DIRTY flag. If
* anything queues a redraw on the actor without specifying an effect
* or with an effect that is lower in the chain of effects than this
* one then that will override this call. In that case this effect
2011-09-12 08:12:14 -04:00
* will instead be called with the %CLUTTER_EFFECT_PAINT_ACTOR_DIRTY
* flag set.
*
* Since: 1.8
*/
void
clutter_effect_queue_repaint (ClutterEffect *effect)
{
ClutterActor *actor;
g_return_if_fail (CLUTTER_IS_EFFECT (effect));
actor = clutter_actor_meta_get_actor (CLUTTER_ACTOR_META (effect));
/* If the effect has no actor then nothing needs to be done */
if (actor != NULL)
_clutter_actor_queue_redraw_full (actor,
0, /* flags */
NULL, /* clip volume */
effect /* effect */);
}