Bug 223477 – net-im/pidgin-2.4.2 randomly crashes on IRC

• Product: Gentoo Linux
• Status: RESOLVED UPSTREAM

[Description Stefan de Konink 2008-05-24 19:20:06 UTC]

Lastnight I have upgraded pidgin, and I noticed different segmentation faults with incomming messages. I see it states that its core is dumped (where?). 

Reproducible: Sometimes

Steps to Reproduce:

[Comment 1 Stefan de Konink 2008-06-26 17:06:29 UTC]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb65146e0 (LWP 24281)]
0xb72f011b in strlen () from /lib/libc.so.6
(gdb) bt
#0  0xb72f011b in strlen () from /lib/libc.so.6
#1  0xb72c156c in vfprintf () from /lib/libc.so.6
#2  0xb72e264c in vasprintf () from /lib/libc.so.6
#3  0xb75b5ff2 in g_vasprintf () from /usr/lib/libglib-2.0.so.0
#4  0x0861cb88 in ?? ()
#5  0x086fe020 in ?? ()
#6  0xbfa35f88 in ?? ()
#7  0xb75a50ce in g_string_append_vprintf () from /usr/lib/libglib-2.0.so.0
#8  0xb76f5c82 in purple_conversation_get_type (conv=0xb7fa6120)
    at conversation.c:565

[Comment 2 Stefan de Konink 2008-06-26 22:40:14 UTC]

(gdb) bt
#0  0xb7fb4410 in __kernel_vsyscall ()
#1  0xb7410533 in __write_nocancel () from /lib/libpthread.so.0
#2  0xb60fbbfc in do_send (irc=0x84c15e8, 
    buf=0x10 <Address 0x10 out of bounds>, len=140729320) at irc.c:95
#3  0xb60fd248 in irc_send (irc=0x84c15e8, buf=0x8556b18 "QUIT :Leaving.\r\n")
    at irc.c:157
#4  0xb60ffcaa in irc_cmd_quit (irc=0x84c15e8, cmd=0xb6104fe6 "quit", 
    target=0x0, args=0x0) at cmds.c:397
#5  0xb60fc313 in irc_close (gc=0x84c1558) at irc.c:471
#6  0xb7715d83 in purple_connection_destroy (gc=0x84c1558) at connection.c:263
#7  0xb7701a22 in purple_account_disconnect (account=0x81799f0)
    at account.c:1174
#8  0xb771585f in purple_connection_disconnect_cb (data=0x81799f0)
    at connection.c:487
#9  0xb75a7d7f in g_timeout_dispatch () from /usr/lib/libglib-2.0.so.0
#10 0x00000001 in ?? ()
#11 0x005a63e6 in ?? ()
#12 0x00000001 in ?? ()
#13 0x00000000 in ?? ()

[Comment 3 Olivier Crete (RETIRED) 2008-06-26 22:47:33 UTC]

This seems like an upstream bug, please file it there

[Comment 4 Stefan de Konink 2008-06-26 22:50:31 UTC]

(In reply to comment #3)
> This seems like an upstream bug, please file it there

I thought the big message always present in the ebuild is to make Gentoo devs file gaim/pidgin bugs upstream.


    ewarn "If you experience problems with pidgin, file them as bugs with"
    ewarn "Gentoo's bugzilla, http://bugs.gentoo.org"

[Comment 5 Olivier Crete (RETIRED) 2008-06-26 23:05:59 UTC]

no, the message is to file it here first... so we can filter bad ones.. Yours clearly seems like an upstream bug (btw, normall you should add your emerge --info here too)

[Comment 6 Stefan de Konink 2008-06-26 23:17:38 UTC]

(In reply to comment #5)
> no, the message is to file it here first... so we can filter bad ones.. Yours
> clearly seems like an upstream bug

Upstream tracker sucks... 
http://developer.pidgin.im/ticket/6187

...and these problems were introduced with 2.4.1 -> 2.4.2. But some other stupid things were introduced with the new pidgin releases, so it is finally time to choose another gtkbased IM client.

> (btw, normall you should add your emerge
> --info here too)

I'm not a user of emerge ;) And really don't see the requirement for emerge --info in relationship to segmentation faults. Now I agree that a bad compiler can cause all this...

[Comment 7 Olivier Crete (RETIRED) 2008-06-27 03:05:21 UTC]

emerge --info (or equivalent) is mostly to see if you've put non-kosher cflags or something else thats strange... or to be able to reproduce the problem (ie know which use flags you have set)