Re: [jaa@cc.jyu.fi: BUG REPORT: Emacs is aborting sporadically by itself]

["Jan D." <jan.h.d@swipnet.se>]

Richard Stallman wrote:
> Can anyone help him investigate this crash?  Maybe talk with him
> by phone and tell him what to type.
> 
> ------- Start of forwarded message -------
> X-Authentication-Warning: tukki.cc.jyu.fi: jaa owned process doing -bs
> Date: Mon, 3 Mar 2003 16:06:43 +0200 (EET)
> From: Jani Averbach <jaa@cc.jyu.fi>
> To: <emacs-pretest-bug@gnu.org>
> X-Virus-Scanned: by amavisd-milter (http://amavis.org/)
> Subject: BUG REPORT: Emacs is aborting sporadically by itself
> Sender: emacs-pretest-bug-bounces+rms=gnu.org@gnu.org
> 
> 
> Please describe exactly what actions triggered the bug
> and the precise symptoms of the bug:
> 
> When I was editing files (LaTeX sources, with global-font-lock t)
> emacs is aborting sporadically by itself. Size of the file is about 25k.
> 
> I can't not describe any pattern for that aborting, it would happen during
> editing, or when I am away from keyboard.

> (gdb) bt
> #0  0x405f00b1 in __kill () at __kill:-1
> #1  0x0810ccd1 in fatal_error_signal (sig=18550) at /data/a/users/jaa/sw/emacs/src/emacs.c:379
> #2  <signal handler called>
> #3  0x405f00b1 in __kill () at __kill:-1
> #4  0x0810cd26 in abort () at /data/a/users/jaa/sw/emacs/src/emacs.c:412
> #5  0x08158890 in mark_object (argptr=0x0) at /data/a/users/jaa/sw/emacs/src/alloc.c:4871
> #6  0x081598ce in mark_maybe_object (obj=1486446028) at /data/a/users/jaa/sw/emacs/src/alloc.c:3440
> #7  0x081574c5 in mark_memory (start=0xbfffec00, end=0xbffff84c) at /data/a/users/jaa/sw/emacs/src/alloc.c:3577
> #8  0x0815751b in mark_stack () at /data/a/users/jaa/sw/emacs/src/alloc.c:3820
> #9  0x08157a7b in Fgarbage_collect () at /data/a/users/jaa/sw/emacs/src/alloc.c:4190
> #10 0x0816c7ae in Ffuncall (nargs=4, args=0xbfffedc0) at /data/a/users/jaa/sw/emacs/src/eval.c:2680
> #11 0x0816c3e8 in call3 (fn=407725308, arg1=405971540, arg2=405971852, arg3=949359756) at /data/a/users/jaa/sw/emacs/src/eval.c:2581
> #12 0x0810199f in x_get_local_selection (selection_symbol=405971540, target_type=405971852, local_request=0)
>     at /data/a/users/jaa/sw/emacs/src/xselect.c:429

This and the other stack trace he sent shows a crash inside 
x_get_local_selection.  That function, and x_handle_selection_request where it 
is called from, has local Lisp_object:s but no GCPRO.  Isn't that needed?  I am 
not sure.

Just a thought.

	Jan D.