[Bug: 21.5-b31] seg fault

[Stephen J. Turnbull]

Don Cohen writes:

 > I was having trouble with the keyboard or mouse somehow inserting
 > random junk in my input stream so was about to look at what was
 > actually being sent.  I had put (recent-keys ) into *scratch* and
 > then did m-x eval-print <enter> when I got the crash.
 > I had done this earlier without incident.
 > My init also contained
 > (set-recent-keys-ring-size 10000)
 > 
 > I have no core file but am prepared to help further.

No core, no help.  Lisp by definition can't crash :-), so the problem
is in the C runtime.  Without a C backtrace, there's nothing we can
do.

 > Fatal error (11).

SIGSEGV, which is a memory management problem.  It may be related to
your input issues (that's the most likely candidate since it's
abnormal) but I'd say the odds are way against it.  It's most likely
some other latent issue.

If you can find a way to reproduce under "gdb ./xemacs", or get a core
file and do

 >   gdb /usr/bin/xemacs core

then we can get a backtrace and try to localize, identify, and fix the
problem.

Don't forget

 > If no core file was produced, enable them (often with `ulimit -c unlimited')
 > in case of future reoccurrence of the crash.

Also, your XEmacs appears to be provided by a distro, and very likely
the symbols have been stripped.  If so, please try to find a debug
package for XEmacs, install the exact same version as your XEmacs
package, and use that symbol file to produce a usable backtrace.
Without the symbols, it is almost certain that we can learn nothing.

I don't use Fedora, so I can't help with the details, but Jerry or
somebody will probably be along soon who can help with installing the
relevant debug package and getting gdb to use the symbol file.