[arch-general] xorg crash with setxkbmap frequently use
Hi folks, X server crashes randomly with the following error: [xkb] BOGUS LENGTH in write keyboard desc, expected 5576, got 5592 Graphic card is Intel G965, driver xf86-video-intel, tried both from [extra] and [testing]. I'm using the Caps_Lock key to change the keyboard layout via a script. I do not want to use "grp:caps_toggle" in xkb options since I need to do something else besides just layout switching. My bash script toggles between layouts running "setxkbmap -layout us" or "setxkbmap -layout ru" commands depending on the current one. Once in a while X server crashes on the toggling. Before I used the "stock" method "grp:caps_toggle", and it never crashed. Google finds several bug reports with the similar error string, but they are not related to layout switching. Thanks for ideas! Sergey.
Sergey Manucharian (2009-05-04 22:42):
Hi folks,
X server crashes randomly with the following error:
[xkb] BOGUS LENGTH in write keyboard desc, expected 5576, got 5592
Graphic card is Intel G965, driver xf86-video-intel, tried both from [extra] and [testing].
I'm using the Caps_Lock key to change the keyboard layout via a script. I do not want to use
"grp:caps_toggle"
in xkb options since I need to do something else besides just layout switching. My bash script toggles between layouts running
"setxkbmap -layout us" or "setxkbmap -layout ru"
commands depending on the current one. Once in a while X server crashes on the toggling. Before I used the "stock" method "grp:caps_toggle", and it never crashed.
Google finds several bug reports with the similar error string, but they are not related to layout switching.
Thanks for ideas!
Sergey.
This is a bug with xkbcomp and it has been reported upstream long ago. There is a redhat/fedora bug report [1] (which has been closed, but the fix didn't work) and a freedesktop.org bug report [2], both assigned to Peter Hutterer, who seems rather busy (and the bug seems rather difficult to track down)... I've seen somebody mention that one could disable the computation of keyboard geometry and the bug would be avoided, but I'm just trying to avoid using setxkbmap for the time being. [1] https://bugzilla.redhat.com/show_bug.cgi?id=469572 [2] http://bugs.freedesktop.org/show_bug.cgi?id=20081 -- -- Rogutės Sparnuotos
On Tue, 5 May 2009 14:57:23 +0300 Rogutės Sparnuotos <rogutes@googlemail.com> wrote:
This is a bug with xkbcomp and it has been reported upstream long ago. There is a redhat/fedora bug report [1] (which has been closed, but the fix didn't work) and a freedesktop.org bug report [2], both assigned to Peter Hutterer, who seems rather busy (and the bug seems rather difficult to track down)...
I've seen somebody mention that one could disable the computation of keyboard geometry and the bug would be avoided, but I'm just trying to avoid using setxkbmap for the time being.
[1] https://bugzilla.redhat.com/show_bug.cgi?id=469572 [2] http://bugs.freedesktop.org/show_bug.cgi?id=20081
Thanks for the useful information, Rogutės! Probably instead of using setxkbmap in shell script I'll write a simple C program using x11 libs (as e.g. xfce xkb plugin does). Cheers, Sergey
On Tue, May 5, 2009 at 3:57 PM, Sergey Manucharian <sergeym@rmico.com> wrote:
[1] https://bugzilla.redhat.com/show_bug.cgi?id=469572 [2] http://bugs.freedesktop.org/show_bug.cgi?id=20081
Thanks for the useful information, Rogutės!
Probably instead of using setxkbmap in shell script I'll write a simple C program using x11 libs (as e.g. xfce xkb plugin does).
This might have been fixed with xorg-xserver commit 525aa17f804d37d1cfcbbf6b8e6cddb45e999b20 [1] (I cannot test it, because it's hard to reproduce here since switching to 64bit). I've filed a bug on [2]. [1] http://cgit.freedesktop.org/xorg/xserver/commit/?id=525aa17f804d37d1cfcbbf6b... [2] http://bugs.archlinux.org/task/14816
participants (2)
-
Rogutės Sparnuotos
-
Sergey Manucharian