Description of problem:

The original bug report is for "firefox" package:

   Bug 308701

I think it might not be firefox problem.
Other packages (xchat, thunderbird) seem to have the same problem, although I
have much less information about them because they crash not so often as firefox
does.

   Bug 282691 (xchat)
   Bug 301051 (thunderbird)

They all (xchat/thunderbird/mozilla) show the same error:
===
The program '...' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadRequest (invalid request code or no such operation)'.
  (Details: serial ... error_code 1 request_code 0 minor_code 0)
...
Program exited with code 01.
No registers.
(gdb)
===

I've been running firefox under GDB, with --sync option, and with breakpoint set
to gdk_x_error(), to get backtraces of crashes. Then, I installed several
debuginfo packages to get more detailed backtrace.

You can see all the stack traces as attachments to the foregoing bug reports.

Unfortunately, I'm not familiar neither with GDB, nor with X Window internals.
That's why I don't know what to do next, to investigate the crashes further.
Could you help me?

I'm not sure this is a libX11 problem.
I just know that all the backtraces end up with "_XReply() _XError()" sequence
before getting to gdk_x_error().
In most cases "_XIDHandler() _XGetMiscCode()" sequence precedes _XReply().
One time, though, I saw "_XAllocIDs() _XFlush()" before _XReply().

I saw all those called from XCreatePixmap(), XCreateGC() functions (libX11).
And also from XRenderCreatePicture() function (libXrender library).

You can get more detailed information from the attachments to all those related
bug reports. And I can provide any additional information you need.

There also are two more bug reports that MIGHT be related to the problem. I
don't have any proof that they are related. Just for you information, then:

   Bug 316521 (clock applet dies)
   Bug 277941 (my first report, now marked as related to libwnck)


Version-Release number of selected component:

   libX11-1.0.3-7.fc6.src.rpm
   libXrender-0.9.1-3.1.src.rpm


How reproducible:

Using firefox/xchat/thunderbird from GDB with --sync for a long time, waiting
for them to crash. Breakpoint on gdk_x_error() to get the backtrace.

There are more detailed descriptions of the way I use to reproduce the bug.
You can find that in the related bug reports.
  

Actual results:

The process (firefox, or thunderbird, or xchat) dies.
All windows (of that process) disappear.
(Bugbuddy doesn't catch the situation. I found no core dumps, as well.)