Commit Graph

73 Commits

Author SHA1 Message Date
Jasper St. Pierre
9dc4c5ce73 xi2: Abort early if we don't have a proper GDK window
This can happen in mutter or other applications that use GDK filters
but don't actually create GDK windows for everything they get events
for.

https://bugzilla.gnome.org/show_bug.cgi?id=689401
2012-12-13 10:51:38 -05:00
Owen W. Taylor
0aa989ae76 GtkPlug: fix handling of key events for different layouts
GtkPlug directly handles X KeyPress/Release events, instead of using
translation in GDK (which expects XI2 events for XI2). When this
was done, the handling of the group was stubbed out and never replaced.

Export gdk_keymap_x11_group_for_state() and gdk_keymap_x11_is_modifier()
so we can fill out the fields correctly.

https://bugzilla.gnome.org/show_bug.cgi?id=675167
2012-04-30 16:35:10 -04:00
Sam Thursfield
2e873e3fdc Fix build with XInput < 2.2
Fixes regression introduced in fa8e6f7857
2012-04-24 12:55:39 +01:00
Cosimo Cecchi
8bb3a2001f xi2: ignore extraneous events for smooth scroll
Since the event will be ignored anyway after it's translated (slave
devices are disabled), don't let it run in the smooth scroll code path,
as it will burn our caches for the actual event we're interested in.

https://bugzilla.gnome.org/show_bug.cgi?id=673644
2012-04-22 13:06:39 -04:00
Cosimo Cecchi
37149668c0 xi2: fix thinko in variable type
https://bugzilla.gnome.org/show_bug.cgi?id=673644
2012-04-22 13:01:08 -04:00
Matthias Clasen
5a41b74c5c XI2: More logging for smooth scrolling
In addition to the delta, log device information.
This was triggered by
https://bugzilla.gnome.org/show_bug.cgi?id=673644
2012-04-22 13:00:47 -04:00
Matthias Clasen
432f0151eb xi2: Remove touchscreen heuristics
It doesn't make sense to consider a device a touchscreen if it
doesn't have any touch classes. Even if it has 'touch' in its
name.
https://bugzilla.gnome.org/show_bug.cgi?id=673440
2012-04-06 20:00:35 -04:00
Matthias Clasen
9d1688680c xi2: Don't reset scroll valuators too often
According to XInput gurus, a motion event without valuators
does not invalidate the stored value, so we should keep it.

https://bugzilla.gnome.org/show_bug.cgi?id=671659
2012-03-08 22:35:43 -05:00
Benjamin Otte
858a654676 x11: Avoid spurious focus events on grabs
We want to avoid handling focus events for the private focus window,
otherwise the keyboard grab taken by for example buttons will cause a
spurious FOCUS_OUT/FOCUS_IN on the toplevel.

The code that did this seems to have been lost in the XI2 transition for
GTK3.

This patch reapplies db4a6040af which was
backed out in 18406b7b04 to give
developers a chance to get their X servers fixed. As we want to get this
bugfix in for 3.4, we need to commit it now.

https://bugzilla.gnome.org/show_bug.cgi?id=657578
2012-03-07 15:35:21 +01:00
Matthias Clasen
3dd5e88c07 xi2: Normalize scroll deltas
XI2 provides us with an increment for each scroll valuator,
and by dividing the delta by the increment, we obtain normalized
values in some abstract 'scroll unit'.

For mouse wheels, the evdev driver reports an increment of -1,
so doing this division fixes the inverted scrolling with wheels
that we've seen recently.
2012-03-04 19:12:27 -05:00
Matthias Clasen
9687ac33b8 xi2: Add some more debug output about scroll devices 2012-03-02 23:02:57 -05:00
Matthias Clasen
482f04c89c xi2: add some debug output for smooth scroll events 2012-03-02 09:57:53 -05:00
Javier Jardón
962a5d8b8a Require XInput2.h in X11 backend
Also remove support for XInput.h
2012-03-02 12:52:12 +00:00
Matthias Clasen
352fdc214a Fix build with XI2.x for x < 2 2012-03-01 16:45:23 -05:00
Carlos Garnacho
013da47a07 gdk,xi2: Ensure scroll valuators are reset on window/device switch
This is in order to avoid scrolling glitches as the device causing
scrolling to happen or the Window receiving the events changes.
2012-03-01 16:29:01 -05:00
Carlos Garnacho
147cdd8465 devicemanager,xi2: Implement smooth scrolling
XInput >= 2.1 allows for implementing smooth scrolling,
reporting the different scrolling axes as valuators.
Any change in those will be reported as GdkEventScroll
events with delta_x/y information.

the older kind of scroll events is still handled, and
emulated in devices able to provide smooth scrolling,
setting _gdk_event_set_pointer_emulated() in that case.
2012-03-01 16:28:58 -05:00
Matthias Clasen
e5f7725fe0 gdk: Add some debug output for touch events and devices 2012-03-01 16:25:20 -05:00
Matthias Clasen
f7b7cc22e6 xi2: Translate touch events
Translate XI_TouchBegin/Update/End to GDK_TOUCH_BEGIN/UPDATE/END
events.

At the same time,
set pointer-emulated flags on button events with XIPointerEmulated
and on touch events emulating the pointer.
2012-03-01 16:25:20 -05:00
Matthias Clasen
734c9056c6 xi2: Use the new device types for touch-capable devices
Any device with a XITouchClassInfo with num_touches > 0
qualifies as multitouch.
2012-03-01 16:19:57 -05:00
Carlos Garnacho
fbb0d8e546 xi2: Add major/minor properties to XI2 device manager
This may be used to turn on/off the features that are added to
new XInput2 revisions.
2012-03-01 16:19:57 -05:00
Javier Jardón
9d0febc9a6 Change FSF Address 2012-02-27 17:06:11 +00:00
Matthias Clasen
18406b7b04 Revert "x11: Avoid spurious focus events on grabs"
This reverts commit db4a6040af.

This is causing bad focus confusion e.g. in gnome-terminal.
2011-12-22 16:03:56 +00:00
Benjamin Otte
db4a6040af x11: Avoid spurious focus events on grabs
We want to avoid handling focus events for the private focus window,
otherwise the keyboard grab taken by for example buttons will cause a
spurious FOCUS_OUT/FOCUS_IN on the toplevel.

The code that did this seems to have been lost in the XI2 transition for
GTK3.

https://bugzilla.gnome.org/show_bug.cgi?id=657578
2011-12-19 16:17:14 +01:00
Benjamin Otte
2ea328dfbc x11: Unify focus handling code
This code was essentially copy-pasted in two locations, so unify them in
the same place.

https://bugzilla.gnome.org/show_bug.cgi?id=657578
2011-12-19 16:17:13 +01:00
Carlos Garnacho
358ecad501 gdk,xi2: Make more resilient about not yet known devices
Dealing with disabled devices may turn into hierarchy/device
changed events on device IDs with no backing GdkDevice yet,
so protect against that. The device attachment will be handled
correctly when the device is enabled later.
2011-12-18 18:57:57 +01:00
Matthias Clasen
9b967ab8af XI2: Push error traps around XIQueryDevice calls
This may help against the crashes at resume time that some
people have been hitting when input devices mysteriously
disappear.
2011-11-30 00:16:43 -05:00
Carlos Garnacho
79ff1b669c x11: Don't create GdkDevices for disabled devices
Fixes bug #658379 - Disabled devices still added to list on startup,
spotted by Bastien Nocera. Do not create GdkDevices for disabled
devices on device manager construction, leading to a confusing initial
state.
2011-09-06 20:04:08 +02:00
Carlos Garnacho
b5f30143ed x11: Skip GenericEvents with no cookie data in the xi2 device manager
https://bugzilla.gnome.org/show_bug.cgi?id=653947

It could happen that a cookie event has been already allocated/freed
in an event filter, as it can't be allocated a second time, all GDK
can do is skipping the event. Spotted by Guillaume Desmottes.
2011-07-15 18:33:44 +02:00
Carlos Garnacho
6aab48ead0 gdk/x11: Add gdk_x11_device_manager_lookup()
This function can be used to find the GdkDevice wrapping
an XInput2 device ID. For core devices, the Virtual Core
Pointer/Keyboard IDs (2/3) may be used.
2011-07-15 18:14:38 +02:00
Matthias Clasen
0c162851ed Add a missing break statement 2011-06-14 21:59:55 -04:00
Matthias Clasen
5254fbacfb Avoid an oob read 2011-06-14 21:56:50 -04:00
Carlos Garnacho
d3274e8e06 xi2: Handle XIM modules that replace events
Fixes Bug 645993 - XIM has wierd behaviors. Some XIM modules
filter every key event, possibly replacing it with their own
one. These events usually have serial=0, so make
GdkDeviceManagerXI2 also listen on these.
2011-06-13 07:16:16 -04:00
Carlos Garnacho
a529568f79 gdk: Get the right filter window for XI2 events
GdkDeviceManagerXI2 now implements GdkEventTranslator::get_window,
retrieving the right window from the contained XIEvent.
2011-05-18 22:34:33 +02:00
Carlos Garnacho
53d77a829a gdk: Move XGetEventData() call to GdkEventSource
This function should be called just once across every possible
event translator and GdkFilterFunc that wants to deal with
any cookie event.
2011-05-18 22:34:32 +02:00
Dan Winship
91b5223001 GdkDeviceManagerXI2: don't set user_time on ButtonRelease
The XI2 device manager was mistakenly setting the window user_time on
both ButtonPress and ButtonRelease, which meant that processes that
tried to launch another process based on the time of a ButtonPress
event would end up always focus-stealing-preventing the new app.

https://bugzilla.gnome.org/show_bug.cgi?id=647275
2011-04-12 12:53:47 -04:00
Carlos Garnacho
db74a6e816 GdkDevice: Remove duplicate private function
_gdk_device_get_axis_use() dates back to pre-sealing, when the
xi2 work began, this remaining can be gone with a public
gdk_device_get_axis_use() function already in place.
2011-04-08 16:39:45 +02:00
Dan Winship
605c383c78 GdkDeviceManagerXI2: process send_event core events
XSendEvent doesn't currently work with XI2 events, so add code to
translate core events when they have the send_event flag.

(We still don't actually select for core pointer/keyboard events, so
we will only receive send_event events that are sent with a 0
event_mask.)

https://bugzilla.gnome.org/show_bug.cgi?id=644847
2011-03-17 17:18:33 -04:00
Carlos Garcia Campos
7aeb3c725b xi2: Ignore button release events of button 4-7
We treat button presses of button 4-7 as scroll wheel events, so release
events should be ignored.

https://bugzilla.gnome.org/show_bug.cgi?id=643925
2011-03-06 12:42:02 +01:00
Matthias Clasen
728a45fc1f Remove confusion over XI2 defines
Some places were using XINPUT_2, others XINPUT_XI2. Since configure
defines XINPUT_2, use that throughout.
2011-02-11 23:30:01 -05:00
Benjamin Otte
de9a082ddb x11: Make headers identical no matter if we run with or without XI2
Previously we weren't installing the device headers when compiling
without XINPUT support. But we would include them from gdkx.h, so
essentially the build was broken.

With this patch the types will exist but not do anything.
2011-02-12 01:27:40 +01:00
Benjamin Otte
30abd4d8b5 x11: Fix struct definition for GdkDeviceManagaerXI2 2011-02-12 00:16:03 +01:00
Benjamin Otte
e9ff457a40 x11: Make the device manager struct actually subclass the correct device 2011-02-11 23:53:14 +01:00
Matthias Clasen
d4c1dbe3d7 Pass group state to _gdk_x11_device_xi2_translate_state
This will be necessary to make groups work with XI2.
2011-02-04 21:11:40 -05:00
Matthias Clasen
2c352f51a0 Share code to fill GdkEventKey.string
This code was duplicated between the core and XI2 device manager
implementations.
2011-02-04 21:11:40 -05:00
Matthias Clasen
9ba009a8eb Remove debug spew 2011-02-04 14:05:22 -05:00
Matthias Clasen
272aa7bc55 XI2: don't translate button releases as scroll events
This fixes scroll-by-2 sympoms seen in combo boxes.
https://bugzilla.gnome.org/show_bug.cgi?id=639846
2011-02-04 12:54:44 -05:00
Matthias Clasen
5441ecd350 Fix a warning 2011-02-01 09:48:32 -05:00
Matthias Clasen
38045e789c Remove a dead member
The client_pointer field in GdkX11DeviceManagerXI2 is not used anywhere.
2011-02-01 09:44:33 -05:00
Matthias Clasen
c79413a1df Make xim work with XI2
We cannot totally ignore core key events, since XFilterEvent generates
those at times and relies on them having an effect.
https://bugzilla.gnome.org/show_bug.cgi?id=640965
2011-02-01 09:42:53 -05:00
Matthias Clasen
454c36523a Silence new gcc warnings
gcc 4.6.0 has started to warn about set-but-unused variables.
So don't do that, then.
2011-01-23 18:50:09 -05:00