Commit Graph

4529 Commits

Author SHA1 Message Date
Emmanuele Bassi
b6eb5728e1 stage: Make per-actor motion event delivery accessors public
Complete the quest of commit bc548dc862
by making the ClutterStage methods for controlling the per-actor motion
and crossing event delivery public, and deprecating the global ones.
2011-06-20 11:41:28 +01:00
Robert Bragg
b66c22ac1c actor: documents _get/apply_relative_transform_matrix
Although this patch doesn't make them public, it documents the
_clutter_actor_get/apply_relative_transform_matrix functions so they
could easily be made public if desired. I think these API could be
useful to have publicly, and I originally documented them because I
thought they would be needed in the MX toolkit.
2011-06-17 18:38:29 +01:00
Neil Roberts
87d55ffcfa actor: Don't use propagated_one_redraw to determine effect dirty state
Previously ClutterActor was using priv->propagated_one_redraw to
determine whether to pass CLUTTER_EFFECT_PAINT_ACTOR_DIRTY to the
paint method of the effect. This isn't a good idea because the
propagated_one_redraw flag is cleared whenever clutter_actor_paint is
called, even if the actor isn't actually painted because of the zero
opacity shortcut. Instead of this, ClutterActor now has a separate
flag called is_dirty that gets set whenever queue_redraw_full is
called or whenever the queue redraw signal is bubbled up from a child
actor. The flag is only cleared in clutter_actor_paint if the effects
are actually run. Therefore it will stay set even if the opacity is
zero or if the parent actor decides not to paint the child.

Previously there were two places set propagated_one_redraw to FALSE -
once if the opacity is zero and once just before we emit the paint
signal.  Now that propagated_one_redraw is only used to determine
whether to pass on the queue redraw signal it seems to make sense to
just clear it in one place right at the start of clutter_actor_paint.

https://bugzilla.gnome.org/show_bug.cgi?id=651784
2011-06-17 18:06:51 +01:00
Robert Bragg
1720b77d29 actor: review use of _apply_modelview_transform_recursive
On reviewing the clutter-actor.c code using
_apply_modelview_transform_recursive I noticed various comments stating
that it will never call the stage's ->apply_transform vfunc to transform
into eye coordinates, but actually looking at the implementation that's
not true. The comments probably got out of sync with an earlier
implementation that had that constraint. This removes the miss-leading
comments and also updates various uses of the api where we were manually
applying the stage->apply_transform.
2011-06-17 17:44:16 +01:00
Robert Bragg
ddc9eb5fa5 group: Handle list modification while destroying children
This makes the dispose code that destroys children resilient against
priv->children being modified during child destruction.
2011-06-17 17:23:32 +01:00
Robert Bragg
1741c805be debug: Use cogl_primitive api for paint-volume visualization
Instead of using the cogl_vertex_buffer API this uses the more concise
cogl_primitive API instead. The aim is to get rid of the
cogl_vertex_buffer API eventually so we should be trying out the
replacement API wherever possible.
2011-06-17 16:59:20 +01:00
Robert Bragg
7f64772984 debug: disable culling with CLUTTER_PAINT=paint-volumes
When using CLUTTER_PAINT=paint-volumes to visualize the paint-volumes of
actors we were already disabling clipped-redraws because we are drawing
extra geometry that the actors don't know about but we didn't disable
culling. This was resulting in actors disappearing while using this
debug option.
2011-06-17 16:58:36 +01:00
Robert Bragg
d7bf214397 debug: avoid visualization code during picking
This makes sure we don't try and draw paint-volumes or culling results
during a pick cycle since that results in us reading back invalid ids
from the pick-buffer.
2011-06-17 16:58:36 +01:00
Neil Roberts
1e2e0f21b8 clutter-actor: Add a debug flag for disabling offscreen redirect
This adds CLUTTER_PAINT=disable-offscreen-redirect to help diagnose
problems with the correct opacity changes. This just makes it so that
it never installs the flatten effect so it will never automatically
redirect an actor offscreen.
2011-06-17 12:15:19 +01:00
Emmanuele Bassi
78fbac71f2 keysyms: Update the key symbols
Re-sync with the X.org headers.
2011-06-17 12:14:15 +01:00
Emmanuele Bassi
c79f613c25 Remove an unused variable
And the compiler warning that comes with it.
2011-06-17 12:03:21 +01:00
Robert Bragg
e936b9591a paint-volumes: remove is_axis_aligned assert for _union
This removes the pv->is_xis_aligned assertion in
_clutter_paint_volume_union. We were already considering the case where
the second volume may not be axis aligned and aligning it into a
temporary variable in that case, but we now also consider that the first
pv may also not be aligned.
2011-06-16 15:28:23 +01:00
Robert Bragg
4dc30c255f paint-volumes: avoid is_complete assert in _axis_align
The removes the pv->is_complete assertion from
_clutter_paint_volume_axis_align() and instead if the volume isn't
complete it calls _clutter_paint_volume_complete().
2011-06-16 15:28:23 +01:00
Robert Bragg
923b1657d9 paint-volumes: fix unioning with empty volumes
When calculating the union of a volume with an empty volume we aim to
simply take the contents of the non-empty volume, but we were not
copying the flags across. We now use
_clutter_paint_volume_set_from_volume which copies all the flags except
the is_static flag.
2011-06-16 12:40:22 +01:00
Robert Bragg
394512b274 paint-volumes: remove alignment constraint for _set_origin
The implementation for clutter_paint_volume_set_origin can cope with
volumes that are not axis aligned so this remove the is_axis_aligned
assertion.
2011-06-16 12:40:22 +01:00
Robert Bragg
30eb3be781 paint-volumes: avoid trashing is_static state
In _clutter_paint_volume_set_from_volume we were using memcpy to simply
copy everything from one volume to another, but that meant we were
trashing the is_static flag which determines if the destination
paint-volume was slice allocated or not.
2011-06-16 12:40:11 +01:00
Robert Bragg
a9789616b9 paint-volumes: remove alignment constraint for completion
This removes the constraint that a paint-volume must be axis aligned
before _clutter_paint_volume_complete can be called. NB: A paint volume
is represented by one origin vertex and then three axis vertices to
define the width, height and depth of the volume. It's straightforward
to use the vectors from the origin to the axis vertices to deduce the
other 4 vertices so we can remove the is_axis_aligned assertion.
2011-06-16 12:39:51 +01:00
Robert Bragg
34600949ed clutter-stage: Allow a wider range of visible z values
Since eef9078f the translation of the camera away from the z=zero
plane was hardcoded at 50 which is approximately half way between the
default z_near and z_far values. This ended up with quite a small
distance in user-space coordinates to the far plane with the default
stage size and this was causing test-texture-quality to clip the actor
early.

This patch makes it try to calculate a reasonable value for the
position of the z=0 plane as well as a value for z_far so we maximize
the space in between the z=0 plane and the near plane and we have a
predictable amount of space behind the stage before hitting the far
clipping plane, while considering the trade off of loosing depth
precision by pushing the far plane too far back relative to the near
plane.

With the default fov of 60° it's not possible to use the stage size to
define the gap in-front of the stage plane; only ~87% of the stage size
is possible as an upper limit. We make 85% of the stage_height available
assuming you have a fov of 60°. We consistently provide 10 times the
stage height of space behind the stage regardless of the fov.

It seems worth noting here that we went around in circles a few times
over how to calculate the gaps since there are a number of trade offs to
consider and they also affect the complexity of the solution. In the end
we went for simplicity but commented the issues well enough hopefully so
we can develop a more elaborate solution if we ever have a use-case.

http://bugzilla.clutter-project.org/show_bug.cgi?id=2625
2011-06-15 18:17:06 +01:00
Neil Roberts
2022e4c100 clutter-stage: Don't override the user's perspective matrix
Since eef9078f ClutterStage updates the aspect ratio of the
perspective matrix whenever the size of the stage changes. This meant
that if an application tries to set its own perspective matrix then
part of it would get overridden. It's not really clear what the
use-case of setting the perspective on the stage should be but it
seems like the safest bet is to always try to preserve the
application's request. The documentation for the function has been
tweaked to discourage its use.

http://bugzilla.clutter-project.org/show_bug.cgi?id=2625
2011-06-15 18:03:53 +01:00
Robert Bragg
d90c849e80 make some parts of the stage-window interface optional
Some parts of the StageWindow interface aren't meaningful for all window
systems. This makes stage_window_set_title/fullscreen/cursor_visible
optional instead of requiring those window systems to implement empty
stubs. Notably the empty stubs we had in the Cogl backend (previously
the EGL backend) used g_warning to report the feature as unsupported and
that was causing conformance test failures.
2011-06-15 15:46:38 +01:00
Emmanuele Bassi
fa8e6d368a Do not include clutter-egl.h unconditionally 2011-06-15 13:05:18 +01:00
Robert Bragg
6a73cece71 backend-cogl: fix s/egl/cogl/ naming left-over 2011-06-15 12:43:47 +01:00
Robert Bragg
b3c93bca20 build: Adds missing clutter-glx.h 2011-06-15 12:41:17 +01:00
Robert Bragg
dd02c56bdb build: Adds missing clutter-egl-headers.h 2011-06-15 12:40:02 +01:00
Robert Bragg
2d8083bab8 unify egl and glx backends as "cogl" backend
Since GLX and EGL are abstracted by Cogl the two backends are both
implementing everything using the Cogl API and they are almost
identical.

This updates the egl backend to support everything that the glx backend
supports. Now that EGL and GLX are abstracted by Cogl, the plan is that
we will squash the clutter-egl/glx backends into one. Since the EGL
backend in clutter can conditionally not depend on X11 we will use the
EGL backend as the starting point of our common backend.

https://bugzilla.gnome.org/show_bug.cgi?id=649826
2011-06-14 20:35:18 +01:00
Robert Bragg
951cb82ec5 egl: fix args to cogl_onscreen_x11_set_foreign_window_xid
We hadn't updated the egl backend inline with a change to the arguments
that cogl_onscreen_x11_set_foreign_window_xid would expect and that was
causing a compilation error.
2011-06-14 20:19:27 +01:00
Emmanuele Bassi
05bcd40729 stage: Ignore do_update() on unrealized stages
An unrealized stage has no resources set yet, so we might as well skip
it during the master clock spin.

https://bugzilla.gnome.org/show_bug.cgi?id=652566
2011-06-14 16:11:54 +01:00
Emmanuele Bassi
81e9bf6127 backend: Do not call ensure_context() with NULL stage
Always call ensure_context() if we have a valid Stage.

https://bugzilla.gnome.org/show_bug.cgi?id=652566
2011-06-14 15:55:06 +01:00
Emmanuele Bassi
0b20739232 glx: Ignore ensuring context on NULL stages
https://bugzilla.gnome.org/show_bug.cgi?id=652566
2011-06-14 15:55:06 +01:00
Emmanuele Bassi
a7fc76a1c5 build: Disable deprecated symbols during development cycles
And add a configure switch to toggle them.
2011-06-14 14:50:10 +01:00
Emmanuele Bassi
4382e6588c glx: Remove unused Atom members from ClutterBackendGLX
The atoms were moved to the shared X11 backend structure ages ago.
2011-06-14 11:36:01 +01:00
Colin Walters
b046bc354c ClutterModifierType: Port fix from GDK to clutter's GDK copy
See https://bugzilla.gnome.org/show_bug.cgi?id=634994 where GDK was
modified to include all flag values, which helps bindings.  Port this
fix to Clutter's copy of GDK.

https://bugzilla.gnome.org/show_bug.cgi?id=650329
2011-06-14 00:05:33 +01:00
Philippe Normand
494629d4b9 clutter-text: new clutter_text_get_layout_offsets API
This new function can be used to retrieve the coordinates where the
ClutterText will draw the PangoLayout representing the text.

http://bugzilla.clutter-project.org/show_bug.cgi?id=2553
2011-06-13 16:15:22 +01:00
Emmanuele Bassi
74a9b71060 effect: Rename RunFlags to PaintFlags
Since run() was replaced and both paint() and pick() use the
enumeration.

https://bugzilla.gnome.org/show_bug.cgi?id=651700
2011-06-13 16:00:45 +01:00
Tomeu Vizoso
0ede622f51 Give a chance to effects for running when picking
Some effects can change the actor's shape and position, so they need
to run when picking.

https://bugzilla.gnome.org/show_bug.cgi?id=651700
2011-06-13 15:45:36 +01:00
Tomeu Vizoso
700c543850 Rename ClutterEffect::run to ClutterEffect::paint
In preparation for adding ClutterEffect::pick

https://bugzilla.gnome.org/show_bug.cgi?id=651700
2011-06-13 15:44:58 +01:00
Emmanuele Bassi
165b2f9599 docs: Fix Clutter API reference 2011-06-13 15:36:07 +01:00
Emmanuele Bassi
cd2b36ed91 Merge branch 'swipe-action'
* swipe-action:
  test-swipe-action: Clean up the test code
  docs: Add the new actions to the API reference
  gesture-action: Remove the multi-device entry points
  swipe-action: Remove the required devices call
  swipe-action: Clean up
  gesture-action: Clean up
  Add ClutterSwipeAction and ClutterGestureAction
2011-06-13 14:39:03 +01:00
Emmanuele Bassi
4f0769d69d docs: Update missing documentation 2011-06-13 13:47:09 +01:00
Emmanuele Bassi
8625f07c3b script: Plug a leak
The SignalInfo structure gained a couple of string fields that should be
deallocated when destroying the structure.
2011-06-13 13:47:09 +01:00
Emmanuele Bassi
a9dd5abfcc script: Allow warping to states
Do not just allow animating states connected to signals: add a "warp"
optional key that ends up calling clutter_state_warp_to_state(). This
is useful for debugging.
2011-06-13 13:47:09 +01:00
Emmanuele Bassi
3017a48228 Skip introspection for GValue ↔ CoglFixed API
It's pointless to introspect an API that it's not meant to be used
through introspection.
2011-06-13 13:47:09 +01:00
Emmanuele Bassi
d9c1a2ebbb script: Use a weak ref to remove the signal emission hook
When an object goes away we should remove the emission hook for the
state transitions, to avoid leaving hooks around.
2011-06-13 13:47:09 +01:00
Emmanuele Bassi
b33973f9f8 script: Rename "state" → "states"
Make it clear that the key used when defining a state-based signal is
linked to the states of a ClutterState.
2011-06-13 13:47:08 +01:00
Emmanuele Bassi
29d7c5a297 script: Allow resolving signals states added from code
Currently, defining states for object signals can only be done by
defining a ClutterState inside the ClutterScript definition. We should
allow creating a (named) ClutterState in code, and associating it to a
ClutterScript instance — and have the Script resolve the "state" field
of a signal definition correctly.
2011-06-13 13:47:08 +01:00
Emmanuele Bassi
dd8cf63a62 script: Allow connecting signal to ClutterState states
One of the uses of a ClutterState state machine along with ClutterScript
is to provide a quick way to transition from state to state in response
to signal emitted on specific instances.

Connecting a real function, in code, to a specific signal does not
improve the ease of use of ClutterScript to define scenes.

By adding a new signal definition to the current one we can have both a
simple way to define application logic in code and in the UI definition
file.

The new syntax is trivial:

  {
    "name" : <signal name>,
    "state" : <state machine script id>,
    "target-state" : <target state>
  }

The ClutterState instance is identified by its script id, and the target
state is resolved at run-time, so it can be defined both in
ClutterScript or in code. Ideally, we should find a way to associate a
default ClutterState instance to the ClutterScript one that parses the
definition; this way we would be able to remove the "state" member, or
even "style" the behaviour of an object by replacing the ClutterState
instance.

The implementation uses a signal emission hook, to avoid knowing the
signal signature; we check the emitter of the signal against the object
that defined the signal, to avoid erroneous state changes.
2011-06-13 13:47:08 +01:00
Robert Bragg
7b4f7cd3e0 build: include <cairo.h> not <cairo/cairo.h>
cairo.h is intended to be included as <cairo.h> not <cairo/cairo.h> as
is the style for clutter.h. If you have installed cairo to a custom
prefix then using cairo/cairo.h can result in unintentional use of the
system cairo headers, or if they aren't installed then it will result in
a failure to find the header.
2011-06-10 17:57:18 +01:00
Emmanuele Bassi
2c791da70a gesture-action: Remove the multi-device entry points
GestureAction supports a single device/touch point. We'll need touch
events supported in Clutter before adding the ability to set required
device/touch points on gestures.

http://bugzilla.clutter-project.org/show_bug.cgi?id=2585
2011-06-10 15:15:15 +01:00
Emmanuele Bassi
f9ce901496 swipe-action: Remove the required devices call
The SwipeAction requires one device/touch point, and that is the default
for GestureAction — if GestureAction supported multiple devices/touch
points.

http://bugzilla.clutter-project.org/show_bug.cgi?id=2585
2011-06-10 15:15:15 +01:00
Emmanuele Bassi
11d961b605 swipe-action: Clean up
Clean up the documentation, and use a type safe GFlags marshaller for
the ::swept signal's @direction argument - instead of a generic integer
marshaller.

http://bugzilla.clutter-project.org/show_bug.cgi?id=2585
2011-06-10 15:15:15 +01:00