[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: CVS emacs cores now and then on Solaris
From: |
Eli Zaretskii |
Subject: |
Re: CVS emacs cores now and then on Solaris |
Date: |
Wed, 29 Oct 2003 20:14:46 +0200 |
> From: Klaus Zeitler <address@hidden>
> Date: 28 Oct 2003 09:01:40 +0100
>
> I think I can now reproduce this crash. It happens when I have
> gobal-autorevert-mode on, then I load a file, that'll get a lot of input
> in the next minutes (e.g. I just run an image that produced
> 10866067 bytes trace info in a log file in about 4 minutes).
> Then I watch emacs constantly reverting the log buffer while it gets input.
> This works fine for about 2 or 3 minutes but then emacs suddenly pops up
> the question:
> umc11.log changed on disk; really edit the buffer? (y, n, r or C-h)
> and then, when I try to cancel this question with a few C-g, emacs crashes.
Hmm, sounds like some GCPRO-related problem?
> Still the backtrace doesn't show much, but xbacktrace shows more (see below).
>
> Program received signal SIGSEGV, Segmentation fault.
> read_filtered_event (no_switch_frame=0, ascii_required=0, error_nonascii=0,
> input_method=0) at /vol/freeware/cvs/emacs/head/emacs/src/lread.c:518
> 518 if (! NILP (delayed_switch_frame))
> (gdb) bt
> #0 read_filtered_event (no_switch_frame=0, ascii_required=0,
> error_nonascii=0, input_method=0)
> at /vol/freeware/cvs/emacs/head/emacs/src/lread.c:518
> Cannot access memory at address 0x38
Hmm, a smashed stack? Why else could it fail to print the full
backtrace, and complain about the address 0x38?
> (gdb) xbacktrace
> "read-char-exclusive"
> "byte-code"
> "ask-user-about-supersession-threat"
> "set-buffer-modified-p"
> "set-buffer-file-coding-system"
> "after-insert-file-set-coding"
> "insert-file-contents"
> "revert-buffer"
> "auto-revert-buffers"
> "apply"
> "byte-code"
> "timer-event-handler"
> "read-char-exclusive"
> "byte-code"
> "ask-user-about-supersession-threat"
> "set-buffer-modified-p"
> "set-buffer-file-coding-system"
> "after-insert-file-set-coding"
> "insert-file-contents"
> "revert-buffer"
> "auto-revert-buffers"
> "apply"
> "byte-code"
> "timer-event-handler"
> "read-char-exclusive"
Looks like some code that calls itself time and again, repeatedly.
Can you see why does this happen?