<<< Date Index >>>     <<< Thread Index >>>

Bug#219499: segfault in different context



-- 
ciao, |
Marco | [2991 seUW2T9fNK7UQ]
--- Begin Message ---
mutt segfaulted today in a context not involving the display problem I
described yesterday. I had just exited my editor (vim) to send a message,
nad mutt segfaulted before I was able to send the email.  I.e., I got a
segfault in between saving and exiting vim and then "send message" screen
in mutt.

(fortunately the message was still in /tmp)

I still don't have any useful information about what precise
circumstances cause this, but it is happening frequently.  
-- 
Adam Kessel
http://bostoncoop.net/adam

Attachment: signature.asc
Description: Digital signature


--- End Message ---
--- Begin Message ---
On Fri, Nov 07, 2003 at 09:53:14PM +0100, Laurent Fousse wrote:
> > Could you generate core files and see in which function it happens?
> > I've upgraded mutt recently and didn't get such behaviour so far.
> Sorry, I didn't see you already provided such traces (but I still
> don't have any segfault; do you have your mutt running for a long
> time?).

It's quite variable, but this morning, for example, it segfaulted in the
first hour or so after it was executed. So it's not necessarily the case
that it is running for a long time when it segfaults.

Note that before this week, I hadn't had a mutt segfault for months. My
usage and/or configuration haven't changed.
-- 
Adam Kessel
http://bostoncoop.net/adam

Attachment: signature.asc
Description: Digital signature


--- End Message ---
--- Begin Message ---

--- End Message ---
On Tue, Nov 11, 2003 at 12:09:39AM +0100, Marco d'Itri wrote:
> On Nov 07, Adam Kessel <adam@xxxxxxxxxxxxxx> wrote:
>  >I hope that helps.
> I'm sorry, but it's useless. You need to build mutt from source and run
> the unstripped binary in gdb:
> 
> apt-get source mutt
> cd mutt-*
> fakeroot ./debian/rules build
> cd build-tree/mutt
> gdb ./mutt

Bingo. I finally got another segfault running the unstripped mutt under
gdb.

Program received signal SIGSEGV, Segmentation fault.
0x080acb76 in unlink_message (old=0x8, cur=0x8179778) at thread.c:457
457         *old = cur->next;

(gdb) backtrace
#0  0x080acb76 in unlink_message (old=0x8, cur=0x8179778) at thread.c:457
#1  0x080ad5a9 in mutt_sort_threads (ctx=0x8118ad0, init=0) at thread.c:811
#2  0x080ab4fa in mutt_sort_headers (ctx=0x8118ad0, init=0) at sort.c:234
#3  0x080692a5 in update_index (menu=0x814b7b8, ctx=0x8118ad0, check=1, 
oldcount=0, index_hint=8) at curs_main.c:326
#4  0x0806f8f4 in mutt_index_menu () at curs_main.c:501
#5  0x08082550 in main (argc=1, argv=0xbffffa34) at main.c:925

Hopefully this is more helpful.  At the moment, I still have this gdb
session open, in case there's something else I can do to help debug.
-- 
Adam Kessel
http://bostoncoop.net/adam

Attachment: signature.asc
Description: Digital signature