22528 Commits

Author SHA1 Message Date
Robert Bragg
0bf128198d configure: improve the error message if gbm missing
This improves the error message in the case where libgbm is missing when
the KMS egl platform has been enabled.

https://bugzilla.gnome.org/show_bug.cgi?id=706808

Reviewed-by: Neil Roberts <neil@linux.intel.com>
2013-11-27 20:38:00 +00:00
Robert Bragg
6436cd073d Declare interface types as void and remove cast macros
This declares the interface types CoglFramebuffer, CoglBuffer,
CoglTexture, CoglMetaTexture and CoglPrimitiveTexture as void when
including the public cogl.h header so that users don't have to use lots
of C type casts between instance types and interface types.

This also removes all of the COGL_XYZ() type cast macros since they do
nothing more than compile time type casting but it's less readable if
you haven't seen that coding pattern before.

Unlike with gobject based apis that use per-type macros for casting and
performing runtime type checking we instead prefer to do our runtime
type checking internally within the front-end public apis when objects
are passed into Cogl. This greatly reduces the verbosity for users of
the api and may help reduce the chance of excessive runtime type
checking that can sometimes be a problem.

(cherry picked from commit 248a76f5eac7e5ae4fb45208577f9a55360812a7)

Since we can't break the 1.x api this version of the patch actually
defines compatible NOP macros within deprecated/cogl-type-casts.h
2013-11-27 19:33:44 +00:00
Owen W. Taylor
9a471cf08d Don't queue redraws when reallocating actor that haven't moved
When support for implicit animation of actor position was added,
the optimization for not queueing when allocating an actor back
to the same location was lost. This optimization is important
since when we are hierarchically allocating down from the top of
the stage we constantly reallocate the actors at the top of the
hierarchy back to the same place.

https://bugzilla.gnome.org/show_bug.cgi?id=719368
2013-11-26 12:35:33 -05:00
Owen W. Taylor
0d0248b31f Bind constraints: Don't force redraws on source relayout
When the source actor potentially changes size, that shouldn't
necessarily result in the target actor being redrawn - it should
be like when a child of a container is reallocated due to changes
in its siblings or parent - it should redraw only to the extent
that it is moved and resized. Privately export an internal function
from clutter-actor.c to allow getting this right.

https://bugzilla.gnome.org/show_bug.cgi?id=719367

Conflicts:
	clutter/clutter-actor.c
2013-11-26 12:35:07 -05:00
Jasper St. Pierre
d44574f738 wayland-surface: Process damage even if the buffer didn't change 2013-11-26 12:32:41 -05:00
Owen W. Taylor
2e85269368 Don't queue redraws when reallocating actor that haven't moved
When support for implicit animation of actor position was added,
the optimization for not queueing when allocating an actor back
to the same location was lost. This optimization is important
since when we are hierarchically allocating down from the top of
the stage we constantly reallocate the actors at the top of the
hierarchy back to the same place.

https://bugzilla.gnome.org/show_bug.cgi?id=719368
2013-11-26 12:30:00 -05:00
Owen W. Taylor
0b536c02f9 Bind constraints: Don't force redraws on source relayout
When the source actor potentially changes size, that shouldn't
necessarily result in the target actor being redrawn - it should
be like when a child of a container is reallocated due to changes
in its siblings or parent - it should redraw only to the extent
that it is moved and resized. Privately export an internal function
from clutter-actor.c to allow getting this right.

https://bugzilla.gnome.org/show_bug.cgi?id=719367
2013-11-26 11:12:12 -05:00
Jasper St. Pierre
a9424255a5 wayland-surface: Add support for subsurfaces
The state for a subsurface isn't double-buffered yet, though...
2013-11-25 18:33:03 -05:00
Jasper St. Pierre
5089a63d76 wayland-surface: Create the surface actor ourselves
Otherwise, we can't rely on a surface_actor being around to add
ourselves to...
2013-11-25 18:29:58 -05:00
Jasper St. Pierre
f9a2c64460 wayland-surface: Don't require a MetaWindow to process damage...
and ClutterActor will clip to the actor bounds, anyway...
2013-11-25 18:15:20 -05:00
Jasper St. Pierre
7841042a85 wayland-surface: Restructure code flow in wl_surface_attach a bit more... 2013-11-25 18:15:20 -05:00
Jasper St. Pierre
ea1b8cdc22 wayland-surface: Fix copy/paste typo 2013-11-25 18:15:19 -05:00
Jasper St. Pierre
52b48cfbef wayland-surface: Return after posting errors back to the client 2013-11-25 18:15:03 -05:00
Jasper St. Pierre
644f3e1275 window: Make setters for shape/input/opaque regions private again
Wayland now pushes these to the surface actor rather than the toplevel
MetaWindow...
2013-11-25 17:25:56 -05:00
Jasper St. Pierre
304005e04f wayland-surface: Replace MetaWindowActor with MetaSurfaceActor
Since subsurfaces won't have toplevel MetaWindowActors, we need to
use MetaSurfaceActor instead. These are embedded in the MetaWindowActor,
just like MetaShapedTexture was (in fact, MetaSurfaceActor now contains
a MetaShapedTexture)
2013-11-25 17:25:55 -05:00
Jasper St. Pierre
025ab35af7 surface-actor: Add setters for set_input_region / set_opaque_region
Subsurfaces won't have a MetaWindow or MetaWindowActor to fetch these from,
so they need to be tied directly to the MetaSurfaceActor.
2013-11-25 17:25:47 -05:00
Jasper St. Pierre
a27744503b shaped-texture: Make Cullable
Make MetaWindowActor chain up to the generic default MetaCullable
implementation, and remove the helper methods for MetaSurfaceActor
and MetaShapedTexture.
2013-11-25 17:25:47 -05:00
Jasper St. Pierre
1011331caf cullable: Use relative actor coordinates for culling
This ensures that nested cullables work fine.
2013-11-25 17:25:47 -05:00
Jasper St. Pierre
0ccef81789 wayland-surface: Rework flow control for wl_surface_commit 2013-11-25 17:25:46 -05:00
Jasper St. Pierre
4780f74a40 wayland-surface: Properly allow setting the cursor to a NULL buffer 2013-11-25 17:18:22 -05:00
Jasper St. Pierre
57866fb267 wayland-surface: Don't rewrite input/opaque regions if the client doesn't attach any 2013-11-25 17:18:22 -05:00
Jasper St. Pierre
d3bc7570d0 window-actor: Update the shape region when the surface actor changes size 2013-11-25 17:18:22 -05:00
Jasper St. Pierre
74e43a4702 cullable: Turn cull_out / reset_culling into a separate interface
Instead of hardcoded knowledge of certain classes in MetaWindowGroup,
create a generic interface that all actors can implement to get parts of
their regions culled out during redraw, without needing any special
knowledge of how to handle a specific actor.

The names now are a bit suspect. MetaBackgroundGroup is a simple
MetaCullable that knows how to cull children, and MetaWindowGroup is the
"toplevel" cullable that computes the initial two regions. A future
cleanup here could be to merge MetaWindowGroup / MetaBackgroundGroup so
that we only have a generic MetaSimpleCullable, and move the "toplevel"
cullability to be a MetaCullableToplevel.

https://bugzilla.gnome.org/show_bug.cgi?id=714706
2013-11-25 15:21:36 -05:00
Jasper St. Pierre
0764b2058a window-actor: Move the operations we need to do when we cull out here
Soon, we'll move this into a generic MetaCullable interface, but for
now, just put hardcoded knowledge in MetaWindowGroup.

https://bugzilla.gnome.org/show_bug.cgi?id=714706
2013-11-25 15:19:08 -05:00
Jasper St. Pierre
7a787d7946 window-group: Decompose culling operations into two methods
This also fixes a bug in the translation of clip_region.

We will add an interface for this soon, so we can recursively cull like this...

https://bugzilla.gnome.org/show_bug.cgi?id=714706
2013-11-25 15:18:09 -05:00
Jasper St. Pierre
eec0f5df47 display: Fix logic for determining whether our focus was successful
In some cases, we can focus the frame window instead of the client
window, so make sure that our checks include that as well.

https://bugzilla.gnome.org/show_bug.cgi?id=715030
2013-11-25 15:14:18 -05:00
Jasper St. Pierre
47144253e4 cullable: Turn cull_out / reset_culling into a separate interface
Instead of hardcoded knowledge of certain classes in MetaWindowGroup,
create a generic interface that all actors can implement to get parts of
their regions culled out during redraw, without needing any special
knowledge of how to handle a specific actor.

The names now are a bit suspect. MetaBackgroundGroup is a simple
MetaCullable that knows how to cull children, and MetaWindowGroup is the
"toplevel" cullable that computes the initial two regions. A future
cleanup here could be to merge MetaWindowGroup / MetaBackgroundGroup so
that we only have a generic MetaSimpleCullable, and move the "toplevel"
cullability to be a MetaCullableToplevel.

https://bugzilla.gnome.org/show_bug.cgi?id=714706
2013-11-25 15:08:22 -05:00
Jasper St. Pierre
d8c66077f0 window-actor: Move the operations we need to do when we cull out here
Soon, we'll move this into a generic MetaCullable interface, but for
now, just put hardcoded knowledge in MetaWindowGroup.

https://bugzilla.gnome.org/show_bug.cgi?id=714706
2013-11-25 15:08:22 -05:00
Jasper St. Pierre
59a01137e1 window-group: Decompose culling operations into two methods
This also fixes a bug in the translation of clip_region.

We will add an interface for this soon, so we can recursively cull like this...

https://bugzilla.gnome.org/show_bug.cgi?id=714706
2013-11-25 15:08:22 -05:00
Jasper St. Pierre
ce3804ee04 display: Use MetaWindow for auto-raise callbacks
This allows us to autoraise Wayland windows... well, except for the
XQueryPointer, but we'll replace that soon.

https://bugzilla.gnome.org/show_bug.cgi?id=715030
2013-11-25 15:08:21 -05:00
Jasper St. Pierre
f0bc53ce5a display: Fix logic for determining whether our focus was successful
In some cases, we can focus the frame window instead of the client
window, so make sure that our checks include that as well.

https://bugzilla.gnome.org/show_bug.cgi?id=715030
2013-11-25 15:08:21 -05:00
Rui Matos
ba3968a822 keybindings: Fix typo in port to Clutter events
We always want to trigger on key press.
2013-11-25 14:15:28 +01:00
Rui Matos
762fa0e116 compositor: Also redirect windows in wayland compositor mode
XWayland was doing this for us but now it isn't anymore.
2013-11-25 12:00:55 +01:00
Jasper St. Pierre
b8938e9d4d window-actor: Fix a copy/paste typo 2013-11-21 17:36:20 -05:00
Jasper St. Pierre
0be57b621b window-actor: Fix a copy/paste typo 2013-11-21 17:35:52 -05:00
Jasper St. Pierre
7d88b3593b surface-actor: Remove unused field 2013-11-21 17:29:41 -05:00
Jasper St. Pierre
7116d9cedb background-group: Fix cast warning 2013-11-21 16:07:51 -05:00
Jasper St. Pierre
66fb86fd0c background-group: Use the Clutter iteration APIs 2013-11-21 16:07:51 -05:00
Jasper St. Pierre
59cfbb07c8 background-group: Remove some unnecessary cruft 2013-11-21 16:07:51 -05:00
Jasper St. Pierre
858db7081a background-group: Fix cast warning 2013-11-21 15:14:40 -05:00
Jasper St. Pierre
5af7f619c8 background-group: Use the Clutter iteration APIs 2013-11-21 14:24:32 -05:00
Jasper St. Pierre
e5e35e5a7f background-group: Remove some unnecessary cruft 2013-11-21 14:24:32 -05:00
Jasper St. Pierre
6d639ac528 window: Support pinging Wayland surfaces as well 2013-11-21 14:24:04 -05:00
Jasper St. Pierre
aa3643cdde Move window pings to MetaWindow
This will make it possible to use on Wayland as well...
2013-11-21 14:24:04 -05:00
Rui Matos
3cd9a70fea device-manager-evdev: Stop using deprecated libevdev API
Fixes compiler warnings with libevdev >= 0.4 and makes use of a new
function to set the clock id instead of doing the ioctl directly.

https://bugzilla.gnome.org/show_bug.cgi?id=712816
2013-11-21 19:34:15 +01:00
Rui Matos
05e6bcc666 device-manager-evdev: Fix a segfault on device removal
Master devices have a NULL sysfs path so use g_strcmp0 to handle them
without crashing.

https://bugzilla.gnome.org/show_bug.cgi?id=712812
2013-11-21 19:34:14 +01:00
Rui Matos
7d8f72a60e device-manager-evdev: Unref devices on removal
https://bugzilla.gnome.org/show_bug.cgi?id=712812
2013-11-21 19:34:14 +01:00
Rui Matos
ce1f8f1dd0 device-manager-xi2: Fix device instances leaking on removal
Don't add an extra reference when adding to the devices hash table. We
already own the initial reference.

https://bugzilla.gnome.org/show_bug.cgi?id=712812
2013-11-21 19:34:14 +01:00
Rui Matos
18b9384e66 input-device: Fix a GArray leak
https://bugzilla.gnome.org/show_bug.cgi?id=712812
2013-11-21 19:34:14 +01:00
Rui Matos
507d8b1cef input-device: Use g_clear_pointer
https://bugzilla.gnome.org/show_bug.cgi?id=712812
2013-11-21 19:34:14 +01:00