On Sun, Feb 21, 1999 at 07:48:49PM +0000, Raimar Falke <[EMAIL PROTECTED]>
wrote:
>
> Hello
>
> I don't know if this is a known problem, but after a wrote a email mutt
> crashes with an segfault in:
>
> #0 mbox_strict_cmp_headers (h1=0x811b7c8, h2=0x149) at mbox.c:516
> 516 if (h1->received != h2->received ||
> (gdb) bt
> #0 mbox_strict_cmp_headers (h1=0x811b7c8, h2=0x149) at mbox.c:516
> #1 0x806b26e in mutt_reopen_mailbox (ctx=0x80b4958, index_hint=0xbffff110)
> at mbox.c:1118
> #2 0x806a4d6 in mbox_check_mailbox (ctx=0x80b4958, index_hint=0xbffff110)
> at mbox.c:634
> #3 0x80700a0 in mx_check_mailbox (ctx=0x80b4958, index_hint=0xbffff110)
> at mx.c:1168
> #4 0x8057058 in mutt_index_menu () at curs_main.c:311
> #5 0x80695e1 in main (argc=2, argv=0xbffffc44) at main.c:681
>
> I usually use mutt with the option -Z together with procmail. Procmail
> changes the mailbox under mutt. This leads sometimes to
> "Can't sync mailbox" but no crashes. I use lock in .procmailrc
These routines should only be called when the mailbox is modified, not
when mail has just been appended to it by procmail.
> It looks like locking doesn't work like it should. Are there ways to
> prevent this?
I think locking is working just fine. However, I have seen this
coredump in the reopen code for a long time, every once in a while. I
had really hoped that someone had fixed it in my absence from mutt-dev,
but it appears not. I'll try to look into this... It may be a hrard
one to trace, though.
-Daniel
--
Daniel Eisenbud
[EMAIL PROTECTED]
