This crash is with 74b40bba327a2e97780e8e3f995f784add2d6231: X starts and hums along until KDE is mostly up (it's working well enough that you can do things in an xterm started first), and then: Backtrace: 0: X(xf86SigHandler+0x82) [0x80c1b32] 1: [0xb7fb1420] 2: X(XkbWriteXKBGeometry+0x9a7) [0x819d327] 3: X(XkbWriteXKBKeymapForNames+0x59c) [0x819bffc] 4: X(XkbDDXLoadKeymapByNames+0x19f) [0x81a6dff] 5: X(ProcXkbGetKbdByName+0x357) [0x817d077] 6: X [0x81851cb] 7: X [0x8148437] 8: X(Dispatch+0x2ce) [0x8086eee] 9: X(main+0x46b) [0x806d6db] 10: /lib/libc.so.6(__libc_start_main+0xef) [0x4afed4bf] Fatal server error: Caught signal 11. Server aborting I speculate that KDE is doing something XKB-related that's tripping a bug somewhere. I can get a better backtrace with variables and so on if needed (this bug report is just thrown off while I do other stuff). (Before it crashes, the keyboard LEDs work again! Yay! :) )
Same bug has been reported in http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=463222 and http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=467058
(In reply to comment #1) > Same bug has been reported in > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=463222 and > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=467058 > Can you confirm if this is still a problem? XKB has seen some cleanup recently, and I wasn't able to reproduce the bug by starting KDE, openoffice and/or the test program in bug 10525 (referenced by http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=467058).
I tried this with xserver master at b9ca7896356f79ee27be5d5aa62052f6984282b0 and was unable to cause X to crash with oocalc or the testcase at #10525, while I was affected by the crash with Debian's xserver-xorg 2:1.4.1~git20080131-1 and 2:1.4.1~git20080131-2.
I also tried both oocalc and the testcase with xserver-xorg 1.4.0.90 (as of dd6b0de38d649617600a8357e576955c9b831328), and neither crashes the X server.
Please disregard the previous tests. As a control, I tried 74b40bba327a2e97780e8e3f995f784add2d6231 (as per the submitter) and couldn't reproduce a crash with either oocalc or the testcase. Even more puzzling, with Debian's 2:1.4.1~git20080118-1's binaries, both oocalc and the testcase caused a crash, very consistently (every time). However, when I tried building from the source for the binary packages, appropriately patched out of debian/patches, (but not through the package build system) using similar configure options as the debian/rules: configure --prefix=/opt/xserver --build=i486-linux-gnu --disable-xprint --disable-static --with-mesa-source=/usr/share/mesa-source --enable-xorg --enable-xtrap --disable-dmx --disable-xsdl --disable-xfake --disable-xfbdev --disable-kdrive-vesa --disable-lbx --enable-freetype --disable-xorgconfig --disable-xorgcfg --enable-xfree86-utils --with-int10=x86emu --disable-type1 CFLAGS="-Wall -g -O2" I was not able to reproduce the crash with either oocalc or the testcase.
Amit, 2:1.4.1~git20080118-1 is a very old package, don't bother looking at it. Please try at least with the latest xorg-server in unstable, which is 2:1.4.1~git20080131-3 as of today.
At the server-1.4-branch tip (dd6b0de38d649617600a8357e576955c9b831328), I see the same failure as ever. Did you want me to check master too? (It's odd that this is not consistently reproducible. Toolchain problem, maybe?)
(In reply to comment #7) > At the server-1.4-branch tip (dd6b0de38d649617600a8357e576955c9b831328), I see > the same failure as ever. > > Did you want me to check master too? if you can find the time, yes please - much appreciated.
> Amit, > 2:1.4.1~git20080118-1 is a very old package, don't bother looking at it. > Please try at least with the latest xorg-server in unstable, which is > 2:1.4.1~git20080131-3 as of today. I'm aware. I used the old package as a control, since it is known to have the issue, to check whether my testing methods were correct. As it turns out, they weren't.
Amit: what's the status of this bug? which version was the newest you tested?
For me at least, this crash has dissolved as mysteriously as it appeared: X 1.4.2 and KDE 3.5.9 do not exhibit it. (I'm not sure which of these changed to fix it.) (Now my only problem is a multiway CPU-chewing nonresponsive bitchfest between kdesktop, xscreensaver, fvwm, konqueror, and X11... but I've seen that before: it's not a new bug.)
> --- Comment #11 from Nix <nix@esperi.org.uk> 2008-06-28 08:01:58 PST --- > For me at least, this crash has dissolved as mysteriously as it appeared: X > 1.4.2 and KDE 3.5.9 do not exhibit it. (I'm not sure which of these changed to > fix it.) calling it fixed then, please reopen if it pops up again.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.