[224] in bug-owl
Re: SIGPIPE from aim
daemon@ATHENA.MIT.EDU (Jeremy Daniel)
Wed Jul 9 18:40:37 2003
To: bug-owl@MIT.EDU
From: Jeremy Daniel <jdaniel@MIT.EDU>
Date: 09 Jul 2003 18:39:49 -0400
In-Reply-To: <bug-owl:222@unknown-discuss-server>
Message-ID: <skeel0z2ukq.fsf@multics.mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
I've narrowed down the bug a bit further by adding debuging code to my
local version. The bug consistantly happens after the 17th request
for idle time. This means that it will only get triggered for long
buddy lists with a lot of people online.
I have been able to keep the bug from triggering by commenting out the
code in owl_aim_get_idle() and recompiling. However I have been
unable to find a proper solution to this bug yet.
Here are the things I've tried so far:
* I tried to make sure the server wasn't requesting a pause.
* I tried to use aim_conn_setlatency(), but unsurprisingly it caused
owl to hang briefly without appearing to fix anything.
* Looking at the faim code doesn't lead to anything obvious that is
limited to 2^4.
I intend to work on the fix further, but without any fresh ideas I
think I am at a dead end for now. If anyone has any useful theories
I'll be happy to test them.
Jeremy
<daemon@ATHENA.MIT.EDU> (Jeremy Daniel) writes:
> Here's the error Chris Shabsin has been getting, it appears to have
> been introduced recently because I only got it when I upgraded to the
> new CVS version. I am guessing that it's related to a conflict
> between idle time code added and buddy groups, but have not been able
> to narrow it down yet. I plan to trace the bug further by backing out
> changes to aim.c one at a time and trying to trigger the bug. Of
> course if it's solved before I get to it I'd be fine with that too.
> :)
>
> Program received signal SIGPIPE, Broken pipe.
> 0xff218ee0 in _so_send () from /usr/lib/libc.so.1
> (gdb) bt
> Reading in symbols for owl.c...done.
> Reading in symbols for aim.c... and list.c...done.
> #0 0xff218ee0 in _so_send () from /usr/lib/libc.so.1
> #1 0x153a14 in aim_send ()
> #2 0x153cec in aim_bstream_send ()
> #3 0x153f3c in sendframe_flap ()
> #4 0x154148 in aim_tx_sendframe ()
> #5 0x154290 in aim_tx_flushqueue ()
> #6 0x44ff4 in owl_aim_process_events () at aim.c:316
> #7 0x146ec in main (argc=0, argv=0xffbeefc0, env=0xffbeefc4) at owl.c:328
------------------------------------------------------------------
Tulgey Wood - tulgeywood@attbi.com Cell Phone: 617-429-8573 /
4 Farragut Ave (617-629-3983) jdaniel-phone@mit.edu
Somerville, MA 02144-1709 AIM: jdanielmit
Zephyr athena users over the web: http://webzephyr.mit.edu/
