[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [leafnode-list] lockup/hang with 1.9.29.rel



Jan Knutar <jknutar@xxxxxx> writes:

> Since I upgraded to .29, I've experienced several hangs where laefnode 
> becomes unresponsive and starts using 100% cpu. Disconnecting from 
> leafnode leaves the process running, sitll using 100% cpu.
>
> This seems to happen fairly often for me, so I started to run strace on 
> the leafnode process today when i started reading news.
>
> If the strace is of any help for someone more knowledgeable, the last 
> bits of it is at: 
> http://fornax.tk/leafnode.hang.strace
>
> I attached to the leafnode process as well and did a backtrace:
> http://fornax.tk/leafnode.hang.gdb

Please try this patch and let me know if it fixes the problem with the
exact article (if you happen to do such a gdb trace again, please
include the output of "print group->name").

I cannot reproduce the problem here because the prerequisite is that the
currently selected news group is part of another news group name on the
XRef: line of the article you mention, and this group name must come
prior to the name we're looking for -- I don't have such articles. I'm
however confident that the patch below fixes the problem, it'll just
need confirmation. If that's the fix, leafnode-1.9.30 is imminent, this
bug is a show-stopper.

diff -u -r1.81 nntpd.c
--- nntpd.c:1.81	2002/09/24 16:04:01	1.81
+++ nntpd.c	2002/11/06 01:31:22
@@ -522,6 +522,7 @@
 			break;
 		    }
 		}
+		p = q;
 	    }
 	    /* if we don't have a localartno, then we need to mark this
 	     * article in a different news group */


-- 
Matthias Andree

http://sourceforge.net/projects/leafnode/   - leafnode-1 tarballs
http://mandree.home.pages.de/leafnode/beta/ - leafnode-2 tarballs

-- 
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list