Bug#430819: xserver-xorg: update breaks X with spurious AIGLX error
J.J.Green wrote:
> Hi Brice
>
>> Before I reassign to fluxbox, please try
>> startx /usr/bin/xterm
>> then run fluxbox from the xterm in there and tell us what happens. If
>> the problem is only fluxbox related, you should get your X and xterm
>> back after fluxbox exited on startup.
>
> Looks like it is: fluxbox fails to start with an
> "Abort" to stdout, and I get back the X and xterm.
> Trace is attached, seems to be a segfault whilst
> reading /tmp/.X11-unix/X0
Good to know, thanks. This socket is used by fluxbox to communicate with
the X server through the X library. So it could be either fluxbox
crashing because it does not use some X calls correctly, or a bug in a X
library. It reminds me of bug #431746 where xinerama not being reported
correctly by the sever causes some clients to crash.
The next step would be to catch this segfault in gdb so that we know
where the crash occurs. Running 'gdb fluxbox' and entering 'bt full'
after the segfault will give you a backtrace (we might need to install
some debugging packages to get a full backtrace, we'll see).
Brice
Reply to: