Commit Graph

5933 Commits

Author SHA1 Message Date
Jasper St. Pierre
a841fff2ac screen: Laters are not GSources
This warns with the new GLib when we destroy the screen.

https://bugzilla.gnome.org/show_bug.cgi?id=711731
2013-11-19 13:52:56 -05:00
Jasper St. Pierre
26aa10a974 compositor: Clean up comment
So it seems that we were intentionally delaying the output window,
although I wonder why...
2013-11-19 13:52:45 -05:00
Jasper St. Pierre
f6144082b1 compositor: Initialize the plugin manager after initializing the output window
Plugins like gnome-shell may want to grab the output window immediately
before bouncing back to the mainloop.
2013-11-19 13:52:13 -05:00
Marc-André Lureau
957513242c xrandr: use "hotplug_mode_update" property
Use the "hotplug_mode_update" connector property indicating that the
screen settings should be updated: get a new preferred mode on hotplug
events to handle dynamic guest resizing (where you resize the host
window and the guest resizes with it).

https://bugzilla.gnome.org/show_bug.cgi?id=711216
2013-11-19 13:50:15 -05:00
Florian Müllner
17fd25e216 doc: Fix file globs
Any files matching the previously used globs are no longer distributed,
breaking distcheck. Match the actual sources in compositor/, core/, meta/
and ui/ instead.
2013-11-19 13:48:40 -05:00
Sindhu S
e91268a250 Add keyboard shortcut for Always on Top, bug 704831 2013-11-19 13:48:35 -05:00
Jasper St. Pierre
82cb4e8267 display: Don't focus the no-focus-window when sending WM_TAKE_FOCUS to a globally active window
Clients like on-screen keyboards try not to take focus when the user clicks
on their window by setting the Input hint to false. However, due to GTK+ and
GDK bugs, the public API for setting the Input hint to false don't remove
WM_TAKE_FOCUS from WM_PROTOCOLS, unintentionally putting them into Globally
Active mode.

These clients also expect that since they don't want to take focus, they want
the focus to remain on the existing window. In this case, for clients like
on-screen keyboards, it's so they can send synthesized keyboard events to the
focused window.

https://bugzilla.gnome.org/show_bug.cgi?id=710296
2013-11-19 13:46:34 -05:00
Rico Tzschichholz
68eb87cc58 build: Make sure built-sources are introspected as needed 2013-11-19 13:46:22 -05:00
Rico Tzschichholz
f3e52d5b18 build: Do not ship generated source-files in tarball 2013-11-19 13:46:21 -05:00
Jasper St. Pierre
2b2b2d3191 stack: Don't take a weak pointer
Since this is stored in an array full of data that will be copied
around, we can't rely on pointer addresses for every item in a stack
not changing.

I don't see any reason that we even have a weak pointer, either.
This code looks safe to me without it.
2013-11-19 13:10:45 -05:00
Jasper St. Pierre
9461c612de wayland-surface: Only spew about set_buffer_scale if it's not 1
... to stop the GTK+ spam on my system.
2013-11-19 13:00:34 -05:00
Jasper St. Pierre
f0280a8868 wayland-surface: Allow setting a NULL buffer to effectively unmap 2013-11-19 12:47:50 -05:00
Jasper St. Pierre
c749f7b6fb Revert "ff"
This reverts commit 74462133ca.

Pushed by accident, if it wasn't obvious...
2013-11-19 11:22:49 -05:00
Jasper St. Pierre
74462133ca ff 2013-11-18 21:35:30 -05:00
Jonas Ådahl
7c45d6594c MetaWindowActor: Use allocation changes signals for size changed signals
When a Wayland compositor, simply rely on the clutter actor allocation
changed signal to sync geometry and emit window actor size changed
signals.

Attaching a wl_buffer to a MetaShapedTexture will signal allocation
changed on the corresponding MetaSurfaceActor, which the MetaWindowActor
is listening to.

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>

https://bugzilla.gnome.org/show_bug.cgi?id=705502
2013-11-18 21:19:01 -05:00
Jonas Ådahl
ea916b6c49 Introduce MetaSurfaceActor for drawing MetaWindowActor content
Instead of having MetaWindowActor only have one single MetaShapedTexture
as actor drawing its content, introduce a new abstract MetaSurfaceActor
that takes care of drawing.

This is one step in the direction to decouple MetaWaylandSurface with a
MetaWindow and MetaWindowActor (except for shell/xdg surfaces) in order
to finally support subsurfaces like features, or any feature where
window is not drawn using a single texture.

The first step, implemented in this patch, is to not have
MetaWindowActor work directly with a shaped texture. There are still
some cases where it simply gets the texture and goes on as before, but
this should be changed by either removing the need of going via
MetaWindowActor or by adding some generic interface to MetaSurfaceActor
that doesn't limit its functionality to one shaped texture.

There should be no visible difference nor after this patch, but
meta_window_actor_get_texture() and meta_surface_actor_get_texture()
should be deprecated when equivalent functionality has been introduced.

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>

https://bugzilla.gnome.org/show_bug.cgi?id=705502
2013-11-18 21:19:00 -05:00
Jasper St. Pierre
a02d734243 window-actor: Move all buffer management and damage correction here
We want ShapedTexture to be a dumb actor that knows how to
pick/paint fairly easily, without any "platform knowledge", so to say...
2013-11-18 21:19:00 -05:00
Jasper St. Pierre
735b736110 shaped-texture: Use the cogl helper methods for fetching shm_buffer contents
Rather than doing the set_region ourselves.
2013-11-18 21:19:00 -05:00
Jasper St. Pierre
bbbb9ac53c display: Make handle_[x]event private 2013-11-18 21:19:00 -05:00
Jasper St. Pierre
8dd97b4998 compositor: Remove outdated comment
XWayland no longer redirects windows by default.
2013-11-18 21:18:59 -05:00
Jasper St. Pierre
73a9082062 cursor-tracker: Make ensure_wayland_cursor take a ref
This fixes a crash in the last commit because that's what I thought
this would do.

It also simplifies callers drastically...
2013-11-18 19:47:34 -05:00
Jasper St. Pierre
4091f5493d cursor-tracker: Force the grab cursor on grab ops
This makes window moving have a 'window move' cursor, and similar.
2013-11-18 19:34:24 -05:00
Jasper St. Pierre
d96b053c9d cursor-tracker: Use a sync methodology for the cursor image
Rather than have MetaWaylandSeat do all the state management itself,
put the split between the root cursor and the window cursor in the
cursor tracker itself. We'll expand this to add a "grab" cursor in
the next commit, allowing us to force a cursor on grab operations.
2013-11-18 19:25:47 -05:00
Jasper St. Pierre
1f569bef76 Fix dragging on move/resize grabs
We need to use the grab_window for incoming events if we have an active grab.
2013-11-18 18:38:24 -05:00
Jasper St. Pierre
1a88176cc0 keyboard: Fix surface focusing
Bad rebase...
2013-11-18 16:52:10 -05:00
Jasper St. Pierre
662c9729bc trivial: window-actor: Fix formatting 2013-11-17 19:16:30 -05:00
Jasper St. Pierre
56a0dd6b2c trivial: shaped-texture: Move set_base_texture to set_cogl_texture
A simple code cleanup.
2013-11-17 19:16:29 -05:00
Jasper St. Pierre
9dc6028b3d trivial: shaped-texture: Fix documentation typo 2013-11-17 19:16:29 -05:00
Jasper St. Pierre
932e913d88 wayland-surface: Fix formatting 2013-11-14 21:49:51 -05:00
Jasper St. Pierre
b72315e27a wayland-surface: Support xdg_surface_resize as well 2013-11-14 21:48:39 -05:00
Jasper St. Pierre
3c7cd1f38c wayland-surface: Use begin_grab_op for move grabs 2013-11-14 21:34:20 -05:00
Jasper St. Pierre
1946c548bf Bump required Clutter version 2013-11-14 14:40:43 -05:00
Jasper St. Pierre
cf181fe109 display: Adapt to new clutter event filter API 2013-11-14 14:34:53 -05:00
Rui Matos
63b9110f93 wayland: Update internal wayland state unconditionally
Things like idle times and the cursor position need to be updated even
if there's a mutter grab in effect.

https://bugzilla.gnome.org/show_bug.cgi?id=712247
2013-11-13 21:55:05 +01:00
Rui Matos
ae44bff0b1 display: Without a grab op let events go through clutter too
We want events to reach clutter no only when there a compositor modal
grab but also when no grab op is in effect.

https://bugzilla.gnome.org/show_bug.cgi?id=712247
2013-11-13 21:55:04 +01:00
Jasper St. Pierre
23ba3e527f wayland: Support alternative focus modes like focus-follows-mouse
Use the existing code for MetaWindow focus-follows-mouse to support this.
2013-11-12 16:53:15 -05:00
Jasper St. Pierre
08df9bf559 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.
2013-11-12 16:53:15 -05:00
Jasper St. Pierre
648639fffe display: Invoke Wayland from the Grand Central Station callback 2013-11-12 16:53:15 -05:00
Jasper St. Pierre
6cc014a941 wayland: Move cursor tracker handling code to meta_wayland_seat_handle_event 2013-11-12 16:53:14 -05:00
Jasper St. Pierre
b7b95123ed wayland: Drop raise on click too
We'll handle this through GCD as well.
2013-11-12 16:53:14 -05:00
Jasper St. Pierre
153463790a wayland: Drop focus on click
The native raise-on-click works with Wayland clients through the standard
Grand Central Station event handling, with MetaWindow.
2013-11-12 16:53:14 -05:00
Jasper St. Pierre
875bbec949 wayland: Drop old hack to synthesize motion events
mutter now uses Clutter events for the source of its drag support, so
we don't need to synthesize X11 events anymore.
2013-11-12 16:53:14 -05:00
Jasper St. Pierre
666e5f1f98 keybindings: Port to Clutter 2013-11-12 16:53:14 -05:00
Neil Roberts
e6790038dd wayland: Use an event filter instead of the captured event signal
In order to see all Clutter events, Mutter was previously installing a
signal handler on the ‘captured-event’ signal on the stage and
additionally using a signal emission hook to cope with grabs. This
changes it to use the new clutter_event_add_filter API so that we can
avoid the signal emission hook hack.
2013-11-12 16:53:14 -05:00
Jasper St. Pierre
e86c53230f display: Move the pointer event handling code to work in terms of Clutter events
There is now a meta_display_handle_event alongside the
meta_display_handle_xevent function which handles events in terms of
Clutter events instead of X events. A Clutter event filter is
registered so that all Clutter events will pass through this function.
The pointer event handling code from the X event version has been moved
into this new function and has been modified to use the details from
the Clutter event instead of the X event. This is a step towards
moving all of the event handling code over to use Clutter events.

Based-heavily-on-a-patch-by: Neil Roberts <neil@linux.intel.com>
2013-11-12 16:53:14 -05:00
Jasper St. Pierre
4ea4658abf display: Split out other event handling as well
This vastly simplifies the code in event_callback and allows us to potentially
take more of this logic and punt it to Wayland-specific backends.
2013-11-12 16:53:14 -05:00
Jasper St. Pierre
58f6ab0a27 display: Split out input event handling
This helps with git's diff code by moving all the giant indentation
changes, letting us know what's going on better when we move to
Clutter for event handling.
2013-11-12 16:53:14 -05:00
Jasper St. Pierre
9c0cc664d1 display: Rename filter_out_event to bypass_gtk
This clarifies the intent of the variable and makes it match bypass_compositor
2013-11-12 16:53:14 -05:00
Jasper St. Pierre
84d26e31f1 display: Move barrier event processing to barrier.c
This puts all the code that's easily ported to Clutter events in one giant
switch statement, as Clutter doesn't translate barrier events.
2013-11-12 16:53:14 -05:00
Jasper St. Pierre
64a848fcb7 display: Add a quick out path to handle_xevent
This makes the code flow more like the scheme we usually use, and allows
us to flatten out the method somewhat.
2013-11-12 16:53:14 -05:00