Re: CVS emacs cores now and then on Solaris

[Eli Zaretskii]

> Date: Fri, 24 Oct 2003 10:03:08 +0200 (MEST)
> From: Klaus Zeitler <address@hidden>
> 
> emacs crashes occasionally. I haven't figured out the reason yet. A few times
> it happened after I hammered (annoyed by my sluggish workstation) a few
> C-g into emacs.
> The backtrace is short for a change :-):
> 
> (gdb) bt
> #0  0xfec1f930 in _libc_kill () from /usr/lib/libc.so.1
> #1  0x000e54c0 in fatal_error_signal (sig=11) at 
> /vol/freeware/cvs/emacs/head/emacs/src/emacs.c:384
> #2  <signal handler called>

This backtrace is useless, since it doesn't show anything before the
signal occured.

> Does anybody have an idea of what's happening? Or can anybody give me a few
> hints how to investigate this?

Does it help to read the file etc/DEBUG in the distro?

I assume that the above backtrace is from a core file.  Running under
GDB to begin with would have provided more info about what Emacs was
doing immediately prior to the crash.

Also, the xbacktrace command supplies the Lisp-level backtrace, but it
only works in a live Emacs session (i.e., again if you run Emacs under
GDB).