Commit Graph

4595 Commits

Author SHA1 Message Date
Carlos Garnacho
029d69919b wayland: Rename touch_handle_cancel_event() to meta_wayland_touch_cancel()
This will serve as a generic call to issue touch cancellation on clients,
useful for the gesture tracker support.

https://bugzilla.gnome.org/show_bug.cgi?id=733631
2014-07-24 00:14:03 +02:00
Carlos Garnacho
5e395fb676 wayland: Fix infinite loop in touch_handle_cancel_event()
https://bugzilla.gnome.org/show_bug.cgi?id=733631
2014-07-24 00:13:26 +02:00
Carlos Garnacho
ac448bd42b seat: Listen for ClutterDeviceManager signals in order to update capabilities
The capability flags are determined from the device types of the slave devices
that are currently attached. This also happens whenever a device is added or
removed, so the capabilities are kept up to date, and clients know about these.

On VT switch, all slave devices are temporarily removed, so the cascade of
signals will make the seat end up with capabililities=0 while input is suspended.

https://bugzilla.gnome.org/show_bug.cgi?id=733563
2014-07-23 22:07:16 +02:00
Rui Matos
f55737ec06 compositor: Sync X drawing only once per frame
We only need to call XSync() once per frame to synchronize X with GL
drawing.

https://bugzilla.gnome.org/show_bug.cgi?id=728464
2014-07-23 13:59:18 -04:00
Carlos Garnacho
32565e096d wayland: Notify modifiers after keymap changes
Anytime the keymap is changed, either directly, or indirectly through the
keyboard capability being released/initialized, there should be a
notification of the modifiers being changed too.

https://bugzilla.gnome.org/show_bug.cgi?id=733563
2014-07-23 14:53:46 +02:00
Carlos Garnacho
1677a068ce wayland: Unset keyboard/pointer focus when releasing the data for these devices
Otherwise the focus_surface_listener list element becomes stale, and then
mangled if the devices' data is initialized again, and the memory memset().

https://bugzilla.gnome.org/show_bug.cgi?id=733563
2014-07-22 18:01:29 +02:00
Carlos Garnacho
a02b8441b1 wayland: set the interface vfuncs when declaring the touch interface
Otherwise the NULL vtable would be accessed when trying to release the touch
device.

https://bugzilla.gnome.org/show_bug.cgi?id=733563
2014-07-22 18:01:29 +02:00
Carlos Garnacho
dfe1c106f2 wayland: set the interface vfuncs when declaring the keyboard interface
Otherwise the NULL vtable would be accessed when trying to release the keyboard.

https://bugzilla.gnome.org/show_bug.cgi?id=733563
2014-07-22 18:01:29 +02:00
Jasper St. Pierre
e49bbe2ed8 cursor-tracker: Don't require a screen
This allows us to do initialization earlier and not have to poke into
Wayland internals from the cursor tracker.
2014-07-22 11:10:14 -04:00
Adel Gadllah
6ee5a1437c x11: Move event filter to backend connection 2014-07-19 18:51:26 +02:00
Jasper St. Pierre
c9a2a561f8 wayland: Use a named #define for the xdg-shell version 2014-07-17 17:48:30 -04:00
Jasper St. Pierre
ac0b2fdc6f wayland: Use our named #define for the wl_shell version 2014-07-17 17:47:14 -04:00
Jasper St. Pierre
46b9984414 wayland: Don't leak an extra xdg_shell resource
No idea how this happened...
2014-07-17 17:47:10 -04:00
Jasper St. Pierre
f9743e2174 wayland: Fix indentation 2014-07-17 17:39:43 -04:00
Jasper St. Pierre
a3c55c2692 wayland: Don't emit SIZE_CHANGED on every commit
Save a little bit of effort here.
2014-07-17 17:27:52 -04:00
Jasper St. Pierre
692eb4d957 wayland: Replace set_margin with set_window_geometry 2014-07-17 16:45:37 -04:00
Jasper St. Pierre
826751429b wayland: Treat the size in the configure event as frame rect coordinates 2014-07-17 16:45:37 -04:00
Jasper St. Pierre
5621d3c0c7 wayland: Fix the sign for the geometry coordinates
I got the spaces for window geometry coordinates vs. surface coordinates
mixed up.
2014-07-17 16:42:41 -04:00
Jasper St. Pierre
b5f46c9171 wayland: Calculate the window geometry based on all subsurfaces
Not just the main surface.
2014-07-17 16:20:19 -04:00
Jasper St. Pierre
166b8c042c wayland: Don't fizzle out attaches 2014-07-17 16:18:01 -04:00
Jasper St. Pierre
d34ece6f1a wayland: Don't send superfluous configures for uninteresting grab ops 2014-07-17 16:03:03 -04:00
Jasper St. Pierre
679844f791 display: Call meta_window_grab_op_ended after everything else is reset
This properly sends the resizing state to the correct Wayland surface.
2014-07-17 16:03:01 -04:00
Jasper St. Pierre
9316256e10 wayland: Use the variables we already have 2014-07-17 15:27:47 -04:00
Jasper St. Pierre
90d7737fc1 wayland: Don't require a new buffer to move windows with dx/dy
This doesn't match what Weston does. I don't know of any apps that this
fixes (we don't have any apps that even use non-zero dx/dy, I don't
think), but this is part of a cleanup for window geometry.
2014-07-17 14:51:23 -04:00
Jasper St. Pierre
567ca15610 xdg-shell: Update to latest 2014-07-17 14:51:12 -04:00
Jasper St. Pierre
558b323485 xdg-shell: Fix typo 2014-07-17 14:50:40 -04:00
Jasper St. Pierre
5df59f75cd restart: Fix pointer/XID confusion
This causes an error in recent gcc.
2014-07-17 11:50:58 -04:00
Adel Gadllah
029673d0ee Revert "backend-x11: Take the touch grab on the stage window"
This reverts commit 3b85e4b2b9.

This breaks touch support; reverting would break wayland
(is what this patch tried to fix; we should find a better solution
that works on both).
2014-07-17 17:18:08 +02:00
Owen W. Taylor
3a57f843d5 Add a framework for restarting the compositor with nice visuals
The current GNOME Shell Alt-F2 restart looks very messy and also
provides no indication to the user what is going on. We need to
restart the compositor to switch in and out of stereo mode, so
add a framework for doing this more cleanly:

Additions:

 meta_restart(): restarts the compositor with a message
 MetaDisplay::show-restart-message: signal the embedding
    shell to show a message
 MetaDisplay::restart: signal the embedding shell to restart
    itself.
 meta_is_restart(): indicates whether the current instance is a
                    restart so we can suppress login animations.

A helper program meta-restart-helper holds the composite overlay
window up during the restart to avoid visual artifacts.

https://bugzilla.gnome.org/show_bug.cgi?id=733026
2014-07-16 18:09:25 -04:00
Owen W. Taylor
182a267f69 Connect to ClutterStage::after-paint with connect_after()
Use connect_after() to accomodate code in GNOME Shell that,
when benchmarking drawing performance, connects to ::after-paint
and calls glFinish(). The timing information from that will be
more accurate if we hold off until that completes before we signal
apps to begin drawing the next frame. If there are no other
connections to ::after-paint, connect() vs. connect_after() doesn't
matter.

https://bugzilla.gnome.org/show_bug.cgi?id=732350
2014-07-16 13:38:15 -04:00
Jasper St. Pierre
bc8328d7c2 screen: Select for events on the guard window on the backend connection
Otherwise, we'll never see them. This fixes the background menu not
working in gnome-shell.
2014-07-15 17:18:37 -04:00
Jasper St. Pierre
12c1a06e60 screen: Don't select for key and motion events on the root window
These aren't necessary, and we'll never see these events anyway, since
we don't process input events on the frontend connection.
2014-07-15 17:18:37 -04:00
Jasper St. Pierre
b11405570a compositor: Don't bother selecting for events on the COW
We already select all the events we need on the stage window instead.
2014-07-15 17:18:37 -04:00
Jasper St. Pierre
3b85e4b2b9 backend-x11: Take the touch grab on the stage window
So we don't get assert fails when touching the screen on mutter nested.
2014-07-14 17:24:31 -04:00
Jasper St. Pierre
c5abf5ddbb window: Fix meta_window_move_resize_internal for the case of a sole ack
When a Wayland window acks our arrangement and we don't really have
anything to modify, we'll pass a sole flag of META_IS_WAYLAND_RESIZE
to meta_window_move_resize_internal using a garbage rect. The existing
code to calculate the new rectangle couldn't really handle this case,
and so the garbage rectangle accidentally got stored. Revamp the flag
checks to be more clear about it.

This fixes the weird positioning issues that sometimes appear when
resizing weston-terminal among others.
2014-07-14 17:07:31 -04:00
Jasper St. Pierre
2a3d4b62a7 keybindings: Fix backend check in change_keygrabs
We really can't do this unless the backend X server is the same as the
frontend X server, as we pass a frontend XID to the backend, which is
only the case when we're not a Wayland compositor.
2014-07-14 15:31:19 -04:00
Jasper St. Pierre
8ff81bcc37 window: Force updating icons when initially loading them
Otherwise, Wayland windows will never get an icon.
2014-07-14 15:28:37 -04:00
Jasper St. Pierre
7159d3bc35 window: Remove the concept of surface_mapped
The implementation was just wrong. We now consider it an error to attach
a NULL buffer to an xdg_surface. Users should destroy the surface properly.
2014-07-14 14:49:58 -04:00
Jasper St. Pierre
941d202938 window: Move X11-specific icon code to MetaWindowX11 2014-07-14 12:15:23 -04:00
Jasper St. Pierre
5ea0cf8bab iconcache: Remove a useless icon_cache_free
All it does is poke a few fields. There is no point to it.
2014-07-14 12:02:58 -04:00
Jasper St. Pierre
df9a5f867f core: Remove useless meta_invalidate_default_icons implementation
This code was supposed to refresh our default icons when the theme
changed, but it actually was a no-op, since the default icons are cached
in a static variable in MetaUI.

I'm not sure the fact that the fallback icons don't update when the
theme changes is an important enough use case to keep working, but I'm
keeping the skeleton function there in case somebody wants to actually
fix it properly.
2014-07-14 12:00:34 -04:00
Jasper St. Pierre
06e4778072 iconcache: Group declarations where they're used 2014-07-14 11:30:03 -04:00
Jasper St. Pierre
92388892b3 iconcache: Remove dead code 2014-07-14 11:27:51 -04:00
Jasper St. Pierre
5d124ac00a iconcache: Remove commented out code 2014-07-14 11:27:51 -04:00
Jasper St. Pierre
a57c3ae105 iconcache: Remove want_fallback hint
It's unused, and if it was ever set, it would cause an assert fail.
2014-07-14 11:27:50 -04:00
Jasper St. Pierre
55354f5c38 Take the touch grab in MetaBackendX11
This means that it won't trigger when we're running as a Wayland
compositor, which will cause a few warnings and possibly more
brokenness.
2014-07-14 09:41:10 -04:00
Jasper St. Pierre
379bb0a77d screen: Don't make a round-trip to fetch a value we know to be empty 2014-07-14 09:40:43 -04:00
Jasper St. Pierre
39357fc242 window-props: Don't make another round-trip to fetch the opaque region
We already have the property value from our property fetching.
2014-07-13 22:51:25 -04:00
Jasper St. Pierre
799de4f0f4 frames: Don't make a round trip to the X server for every MotionNotify
Instead, simply use the coordinates that were passed to us from the
MotionNotify event.
2014-07-13 17:34:08 -04:00
Jasper St. Pierre
c34b5c1cf0 Makefile: Disable Clutter / Cogl deprecation warnings
These just spam up our logs, and they aren't getting fixed any time
soon for a variety of reasons. Just disable them to ensure that really
important warnings get noticed and fixed.
2014-07-13 11:46:13 -04:00
Jasper St. Pierre
46361c3e28 events: Return early if we close the display
This is so we won't poke into the MetaDisplay, which is invalid memory,
and crash. This can sometimes work right now because GSlice might not
deallocate the object immediately, but it's still not a fun thing to do.
2014-07-13 11:43:28 -04:00
Jasper St. Pierre
f9d33b2efc display: Add an extra check inside meta_display_close
For clarity of what the expected input to this function is.
2014-07-13 11:37:31 -04:00
Jasper St. Pierre
2185904fcb display: Use a local here instead of the global the_display
This will make it easier to split some of this functionality out into
other subfunctions.
2014-07-13 11:35:38 -04:00
Jasper St. Pierre
eb952819c2 monitor: Don't use wl_output_transform in our APIs
Invent our own API so we can build without Wayland headers.
2014-07-13 11:35:01 -04:00
Jasper St. Pierre
d7d8c92a9a Remove unused Wayland includes 2014-07-13 11:35:01 -04:00
Owen W. Taylor
88a7790d44 MetaDisplay: remove meta_display_grab()/ungrab()
Now that we have two connections to the X server, the idea of a
ref-counted server grab that might be held across extended portions
of code is very dangerous since we might try to use the backend
connection while the frontend connection is grabbed.

Replace the only usage (which was local) with direct
XGrabServer/XUngrabServer usage and remove the meta_display_grab()
API.

https://bugzilla.gnome.org/show_bug.cgi?id=733068
2014-07-11 13:39:07 -04:00
Owen W. Taylor
806dabe2d7 Don't grab the server when unmanaging the screen
There's no obvious reason for grabbing the X server when unmanaging
a screen - the only race conditions a server grab solves are those
related to querying and then acting on the results of the query.

Our shutdown sequence is correctly ordered according to the ICCCM -
we first unselect on the root window, and then we destroy the
window owning WM_S<n> so removing the grab should not cause any
problems when we are being replaced with another window manager.

https://bugzilla.gnome.org/show_bug.cgi?id=733068
2014-07-11 13:39:07 -04:00
Jasper St. Pierre
740e7ddd69 window: Manage after setting the OR special case properties 2014-07-10 17:08:14 -04:00
Jasper St. Pierre
bee97ab6a7 window-props: Make sure that recalc_window_type is always called 2014-07-10 17:07:52 -04:00
Jasper St. Pierre
59fc17c355 util: Turn STACK debugging back off
Snuck in by accident.
2014-07-10 16:54:48 -04:00
Jasper St. Pierre
5f25ca0c3a stack-tracker: More small cleanups 2014-07-10 16:29:35 -04:00
Jasper St. Pierre
41b4a079c7 stack-tracker: Remove unused variable 2014-07-10 16:29:35 -04:00
Jasper St. Pierre
e4235376d8 stack-tracker: Add some more debug output 2014-07-10 16:29:34 -04:00
Jasper St. Pierre
043a201f90 stack-tracker: Clean up dumping 2014-07-10 16:29:18 -04:00
Jasper St. Pierre
d1a588a94f stack-tracker: Fix the case where the serial is too old
The code here before was just flat out wrong, and made us consistently
requery the server at startup.
2014-07-10 16:02:01 -04:00
Jasper St. Pierre
01b6d9bfe2 stack-tracker: Fix code style 2014-07-10 16:00:38 -04:00
Jasper St. Pierre
9711d95996 stack-tracker: Fix comments 2014-07-10 16:00:38 -04:00
Jasper St. Pierre
b87db00fdb window-props: Fizzle out PropertyNotifies that happen after initial map
Properties like _NET_WM_DESKTOP and _NET_WM_STATE are supposed to be
ignored after the initial map of the window, so ignore any
PropertyNotifies for these.
2014-07-10 15:05:42 -04:00
Jasper St. Pierre
6190b65056 window-props: Use a flags-based approach for our prop hooks
This allows us to extend the table better with newer flags.
2014-07-10 15:05:42 -04:00
Jasper St. Pierre
10fe91d704 window-props: Don't load WM_STATE
Since we don't do anything with it.
2014-07-10 15:05:42 -04:00
Jasper St. Pierre
36bb05b10e window-props: Fix assert testing
I have *no* idea how this even compiled in the first place. What the hell?
2014-07-10 15:05:42 -04:00
Jasper St. Pierre
a7b7213017 window-props: Use the standard property type system for _NET_WM_WINDOW_TYPE 2014-07-10 15:05:40 -04:00
Jasper St. Pierre
031154a400 window-props: Use the standard property type system for WM_WINDOW_ROLE
There seems to be no reason we can't just use this.
2014-07-10 14:56:23 -04:00
Jasper St. Pierre
463b50e746 screen: Fetch the COW window up-front
XGetCompositeOverlayWindow is a round trip to the X server. We shouldn't
be doing it on every MapWindow.
2014-07-10 14:18:02 -04:00
Jasper St. Pierre
1749a85e96 frame: Only XSync when we're on the X11 compositor backend
The only case we have is the case where the two X11 connections are the
same. When on Wayland, the XSync is costly and expensive, and we should
minimize it.
2014-07-10 13:31:26 -04:00
Jasper St. Pierre
1af9e92e91 Don't add in our old input event mask when we select input events
This causes a round-trip to the server, and just isn't necessary, since
we're always in charge of our event mask.
2014-07-10 13:30:05 -04:00
Jasper St. Pierre
12d0c98fb5 wayland: Remove an unused field from MetaWaylandFrameCallback 2014-07-10 13:30:05 -04:00
Jasper St. Pierre
817c6986a0 wayland: Remove a few unused fields from MetaWaylandCompositor 2014-07-10 13:30:05 -04:00
Jasper St. Pierre
7c68aaea4c data-device: Put the DataDevice implementation in its own struct 2014-07-10 13:30:05 -04:00
Jasper St. Pierre
6fea46e0cf data-device: Use a better variable name than "resource" 2014-07-10 13:30:05 -04:00
Jasper St. Pierre
59fa74fed1 data-device: Don't look up the same client twice in two different ways 2014-07-10 13:30:05 -04:00
Jasper St. Pierre
581335fbea data-device: Start a cleanup 2014-07-10 13:30:04 -04:00
Jasper St. Pierre
9b5d6cc008 wayland: Remove seat->display
It's unused.
2014-07-10 13:30:04 -04:00
Jasper St. Pierre
d0f98ec1ba wayland: Fix include ordering
This puts it in the same order as the types in the struct.
2014-07-10 13:30:04 -04:00
Jasper St. Pierre
a977fcf3d0 wayland: Add a wrapper for set_input_focus
So we're not poking into seat internals.
2014-07-10 13:30:04 -04:00
Jasper St. Pierre
664f6ef420 wayland: Move DataOffer / DataSource into data-device internals 2014-07-10 13:30:04 -04:00
Florian Müllner
06a31992e3 screen: Tweak workspace initialization (again)
Commit 8100cefd4c fixed a crash during workspace initialization by
tweaking the startup sequence; as a result, the plugin (like gnome-shell)
is now started before workspaces are fully initialized, which breaks
some reasonable assumptions (like always having an active workspace).
This is particularly problematic considering that the code making those
assumptions is not necessarily our own (extensions!), so return to
fully initialize workspaces before the compositor again.
At the same time, make sure to only call meta_workspace_activate()
once during initialization to avoid reintroducing the crash.

https://bugzilla.gnome.org/show_bug.cgi?id=732695
2014-07-10 17:50:59 +02:00
Jasper St. Pierre
4e2092d593 gesture-tracker: Rename files to include the meta- prefix 2014-07-08 13:30:26 -04:00
Jasper St. Pierre
9755f308e5 gesture-tracker: Update documentation format 2014-07-08 13:28:48 -04:00
Jasper St. Pierre
09610daea1 gesture-tracker: Don't pass an autodeny-timeout
Just use the default value of 250.
2014-07-08 13:27:38 -04:00
Jasper St. Pierre
13b7fa4eba ui: Bring back the is_input_event helper
This makes the logic clearer and easier to reason about.
2014-07-08 12:59:35 -04:00
Jasper St. Pierre
909a46087f ui: Fix the check for the XInput opcode
When converting is_input_event to be the opposite, I forgot to make this
change.
2014-07-08 12:58:43 -04:00
Jasper St. Pierre
c10c8649f5 window: Reformat a comment 2014-07-08 12:45:34 -04:00
Jasper St. Pierre
82dce6758a window: Remove pending pings in unmanage
This makes sure that we see them for Wayland clients as well, and don't
time out and crash when we're accessing an invalid window / surface.

Spotted-by: Rui Matos <tiagomatos@gmail.com>
2014-07-08 12:17:08 -04:00
Jasper St. Pierre
024652bfb4 window: Set the surface to NULL when unmanaging 2014-07-08 12:14:20 -04:00
Carlos Garnacho
789608b637 gesture-tracker: Implement threshold-based sequence rejection
If a sequence moves past a certain distance without being used by a
gesture, reject it so clients may see and react to it ASAP. This makes
gestures to be began by initially quasi-static touchpoints, in addition to
quasi-simultaneous.
2014-07-08 11:30:12 -04:00
Carlos Garnacho
d05b750b8d frames: Keep information about the ongoing grab operation, and retry if needed.
When a passive touch grab is rejected over the frame, management is punted to
the frame itself, and pointer events emulated, but the attempt to transfer the
grab from the GDK connection to the Clutter one fails with AlreadyGrabbed, and
will fail until the Clutter connection receives the XI_TouchEnd resulting from
XIRejectTouch, gotten after the XI_ButtonPress on the GDK connection.

In order to bypass this shortcoming, store the current grab operation on the
frame as long as the button is pressed, so it is retried once on the next
motion event happening during frame dragging, that will have a recent enough
timestamp to succeed. If no grabbing succeeded, the current grab operation
data will be reset on GDK_BUTTON_RELEASE.
2014-07-08 11:21:36 -04:00
Carlos Garnacho
704cae1de3 backend: Use the most recent event time on replayed touch events
When a touch sequence is passively grabbed and later rejected, events
will be replayed on the next client in propagation order, although those
events (either transformed to pointer events or not) will contain the
original timestamps, this will make grabs fail with InvalidTime if triggered
from the replayed ButtonPress/TouchBegin handler.

In order to work around this, store the most recent event time (presumably
gotten from the XI_TouchEnd caused by the passive grab being rejected), and
use that time on the events being replayed afterwards and grabs, so we don't
possibly fail with InvalidTime if those events result in a compositor grab.
2014-07-08 11:21:29 -04:00
Carlos Garnacho
23b074481a display: Set an X11 passive touch grab on the root window
Touch events will be caught first by the compositor this way,
whenever the MetaGestureTracker notifies of the accepted/rejected
state of a sequence, XIAllowTouchEvents() will be called on it
accordingly, so it is handled exclusively by the compositor or
punted to clients.
2014-07-08 11:21:29 -04:00
Carlos Garnacho
5dc8fa6690 events: Hook MetaGestureTracker to display event processing
Events aren't actually consumed by the MetaGestureTracker, but it
rather defines whether the event will reach clients, or the stage.
2014-07-08 11:21:29 -04:00