7b6400a1b8
glx: Fix drawable private leak on destroy
...
When a drawable goes away, we don't destroy the GLX drawable in full,
since it may be current for a context. This means that when the drawable
is destroyed in full later, the backend doesn't get a chance to
destroy resources associated with the drawable (the DRI2Drawable).
With this patch, we destroy the GLX drawable in full when it goes away
and then track down all contexts that reference it and NULL their
pointers.
2009-04-13 13:17:53 -04:00
140463a197
XQuartz: Use correct values for ProximityIn and ProximityOut
...
This was the other underlying cause of teh 100% CPU tablet issue.
(cherry picked from commit a9cecf34c23fbcd59b56b380c51d31a9fabc3eb7)
2009-04-11 17:26:27 -07:00
80a7bb2605
XQuartz: Solve the tablet 100% CPU bug
...
This happened because we put a byte in the fd to wake up dispatch, but we never actually enqueued anything in mieq because the num_events was 0.
(cherry picked from commit c21ca7558d2faf93c61f5feaafd7c878e9e21942)
2009-04-11 17:26:07 -07:00
6f8f7c78f1
XQuartz: Re-enable background window checking
...
since that code was not the culprit for the wacom tablet, background 100% CPU bug
(cherry picked from commit fc1dc5d71b2a488a8a94d953dd8e67353161a590)
2009-04-11 17:23:03 -07:00
8522a759c9
XQuartz: Only set MotionNotify on activation if it is updated.
...
(cherry picked from commit ae8077a251ef27381a755d57ff974767bda16148)
2009-04-11 17:22:50 -07:00
5e55becddc
XQuartz: Make sure the Fn doesn't trigger unneccessary calls to DarwinUpdateModKeys()
...
(cherry picked from commit 70a18558c6b0a02b633fd8974f002cdf3cdc713e)
2009-04-11 17:22:38 -07:00
bd1d917909
XQuartz: Re-enable Fn as an option for 3button mouse simulation.
...
Patch from Martin Otte <otte@duke.edu>
(cherry picked from commit b5ec3be6b5449b5d575bc1472fdd1c9cb15cb8be)
2009-04-11 17:22:20 -07:00
1b5758bef0
exa: implement UTS based upload through CopyArea
...
- Some image viewers (eog, gqview) trigger the CopyArea path of Xext/shm.c
- I'm not aware of any code path that wouldn't like UTS and trigger this code.
- miDoCopy should handle src coordinate clipping.
- Overlapping blits are obviously not an issue (both would have to be offscreen or not).
2009-04-10 21:14:47 +02:00
567cf67959
DRI2: Synchronize the contents of the real and fake front-buffers
...
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
2009-04-10 12:00:36 -07:00
f1a995d149
DRI2: Do not send the real front buffer of a window to the client
...
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
2009-04-10 12:00:36 -07:00
aa2928325f
DRI2: Add fake front-buffer to request list for windows
...
If a front-buffer is requested for a window, add the fake front-buffer
to the list of requested buffers.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
2009-04-10 12:00:35 -07:00
808fd2c67f
XQuartz: xprSetWindowLevel updated to store the level requested by the WM
...
(cherry picked from commit c28c2ddc3a8f3c5b9beec396953bb3ac9ee4714b)
2009-04-10 00:48:35 -07:00
dfb0d7aefb
XQuartz: Update window levels when changing rootless state
...
(cherry picked from commit 1359ded5bfc14a80fb998b01a54ecacb96c4ff88)
2009-04-10 00:47:40 -07:00
d5ef88d754
XQuartz: Fix window levels for rooted mode to allow showing the menu bar.
...
(cherry picked from commit 80759a4186bf0335edc85aecea2faf11fe09f491)
2009-04-10 00:47:28 -07:00
15146b8637
XQuartz: Properly set the window level for the root window
...
(cherry picked from commit bdf9286d1cbfeaaf8eaf03d28091e91ee587ee25)
2009-04-10 00:47:16 -07:00
8d5dcfe215
XQuartz: Properly set the menu bar and hotkey state when changing rootless mode.
...
Currently no code path exhibits the broken behavior since we only toggle into rootless if we don't have the root.
(cherry picked from commit 970f100ca3c5fc0662ae7658d49d118fbd9de943)
2009-04-10 00:47:03 -07:00
82d7cf5cdc
XQuartz: In rooted mode, make sure we start in the hidden state.
...
(cherry picked from commit 5ecc497f71c2133f773f6c56ad76cb778862ddd6)
2009-04-10 00:46:48 -07:00
e86f4e9302
XQuartz: Send MotionNotify before button presses when X11 is in the background
...
(cherry picked from commit c80d0ec18ef5b842447d31360406d0b5b9424222)
2009-04-10 00:46:35 -07:00
5ccfad8df0
XQuartz: Revert most of the previous override redirect patch
...
The changes actually caused all windows to move to the current space. Instead, we're going with a fix entirely within Xplugin that depends on quartz-wm being the window-manager for now.
(cherry picked from commit 997b6f3142c622541bb5bac98652abae75d1101d)
2009-04-10 00:46:25 -07:00
ff7f019bbc
dix: correctly utilize tracker buffer and protect from timer overruns
...
two small related fixes hard to split up
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-10 13:49:51 +10:00
06aebecb19
dix: fix pointer accelerations remainder handling
...
This didn't really work as intended, but did amazingly well thanks
to roundf() hiding the defect. Cheers!
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-10 13:49:51 +10:00
98f4179156
Use RTLD_DI_SETSIGNAL to catch runtime dynamic loader errors and clean up
...
Based on fix for Sun bug 6813925: Xorg needs to catch ld.so.1 failure
so it can close down devices cleanly
<http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6813925 >
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
2009-04-09 17:10:12 -07:00
a0b6a363dc
Lift fatal signal handlers from DDX'es up to a common DIX implementation
...
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
2009-04-09 17:10:12 -07:00
fcc19e673e
mi: fix wrong (*EnqueueEvent) declaration in miPointerScreenFuncRec.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-09 20:50:32 +10:00
4fee979d06
Xi: silence compiler warnings about "wrong" event types.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-09 20:50:32 +10:00
6b467bf879
xfree86: shut up compiler warnings - typecast to InternalEvent
...
Reported-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-09 20:50:32 +10:00
8a2a184da7
xfree86: fix use of uninitialized variable in DGAProcessPointerEvent.
...
Reported-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-09 20:50:32 +10:00
011cee3103
dix: fix dev/keybd variable mixup.
...
Reported-by: Eric Anhold <eric@anholt.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-09 20:50:32 +10:00
630a6e9d14
mi: fix compiler warning - explicitly typecast to InternalEvent.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-09 20:50:32 +10:00
e3bb7dea06
mi: add prototype for CopyGetMasterEvent.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-09 20:50:32 +10:00
346e71525f
EXA: If the driver can't composite to an a8 mask, try an argb mask for glyphs.
...
Signed-off-by: Michel Dänzer <daenzer@vmware.com>
2009-04-09 09:36:41 +02:00
682d7b5569
xselinux: Don't BadAlloc in List* requests if there are no items to list.
2009-04-09 02:48:04 -04:00
3a0ee199dc
config: fix crash caused by strdup(NULL)
2009-04-09 02:29:28 -04:00
7b3982eb65
glx: Test the error value, not its address...
2009-04-09 08:21:09 +02:00
03aebed519
Use a #define instead of a magic number
...
The number of buffers is likely to change in the future, so having
this as a define is the right way to go.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
2009-04-08 15:10:21 -07:00
0d9f3ca7ea
Allow GLX sources to build against Mesa 7.4 sources
...
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
2009-04-08 15:10:21 -07:00
e8b324102f
xselinux: Don't require incoming context strings to be null-terminated.
2009-04-08 15:10:16 -04:00
df27b870a8
Convert remaining GLX LookupIDByType() calls
2009-04-07 16:28:26 -04:00
92562747a0
Add validGlxDrawable() and use dixLookupResourceByType().
...
Fixes deprecation warnings, and fixes a couple of GLX error codes
for failing drawable lookups.
2009-04-07 16:28:26 -04:00
f70cfc8f90
Don't stomp on dixLookupDrawable() return value in DoCreateGLXPixmap().
2009-04-07 16:28:26 -04:00
30d81ad72e
Make GLX context lookup use dixLookupResourceByType()
2009-04-07 16:28:26 -04:00
66539cc05d
Don't leak default font path when appending built-ins
...
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
2009-04-07 10:08:56 -07:00
0cfd481308
dix: Dont change the keyboard mapping on non-keyboard devices.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-04-07 19:25:17 +10:00
69e73e5ce0
config: Remove useless xf86conf{{m,c,re}alloc},free} macros
2009-04-06 17:56:49 -04:00
485946120a
config: s/xf86configStrdup/strdup/
2009-04-06 17:51:35 -04:00
5f769b7344
kdrive: Fix segfault in tslib support
2009-04-06 13:01:08 -04:00
843166b033
os: signal handlers return void.
2009-04-06 11:05:17 -04:00
aa6fbc2a36
Remove some OS/2 leftovers.
2009-04-06 10:58:47 -04:00
2c03a63ff6
XQuartz: Use updated Xplugin API to send overide-redirect windows to the current space when they're ordered in
...
This fixes the annoying "using a menu moves me to another space" bug
(cherry picked from commit 9e6dab89e08a26d764ba1aeaeb804c3d25c667da)
2009-04-05 16:32:12 -07:00
67621f3b39
XQuartz: Send a MotionNotify event for the mouse cursor when activating X11.app
...
Otherwise if X11.app was activated with a mouse click, the location of the even is the last location of the cursor before X11 was deactivated
(cherry picked from commit c7457d7b31ddd2ddfd04dd6294bed4532664e1ab)
2009-04-05 16:32:02 -07:00
Kristian Høgsberg
Jeremy Huddleston
Maarten Maathuis
Ian Romanick
Simon Thum
Alan Coopersmith
Peter Hutterer
Michel Dänzer
Eamon Walsh
Adam Jackson
David Jander