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

Re: [leafnode-list] timeout connecting to (null) + Segmentation

Matthias Andree wrote:
> Ralf Wildenhues <wildenhu@xxxxxxxxxxxxxxx> writes:
> > connecting to news.cis.dfn.de:nntp: cannot connect: Network is unreachable
> > fetchnews: 0 articles fetched, 0 killed, in 0 seconds
> > fetchnews: Fixing XOVER
> > findgroup("de.comp.os.unix.programming") called without prior readactive()
> Could you provide a backtrace for the crash?

Here we go:  IMHO this happens:  because of forceactive==1
rereadactive() is not called in fetchnews.c main().  Because no servers
are available (in my case, don't know about Volker's) active is not
initialized.  The forked process in fetchnews.c main() tries to fixxover
which barfs because active==NULL.  But see for yourself:

$ gdb sbin/fetchnews ./var/spool/news/interesting.groups/core
(gdb) bt
#0  0x4006abf1 in __kill () frcm /lib/libc.so.6
#1  0x4006a7ff in raise (sig=6) at ../sysdeps/posix/raise.c:27
#2  0x4006c033 in abort () at ../sysdeps/generic/abort.c:88
#3  0x804dfc9 in findgroup (name=0x806a65f "de.comm.software.newsreader") at activutil.c:183
#4  0x805a808 in fixxover () at xoverutil.c:621
#5  0x804da64 in main (argc=3, argv=0xbffff5e4) at fetchnews.c:1749
(gdb) p active
$1 = (struct newsgroup *) 0x0
(gdb) p forceactive
$2 = 1

What else to look at?  I did not reproduce this with fetchnews run
inside gdb because of the child process (parents then finishes without
problems).  Doesn't happen when I don't specify -f on the command line.

BTW: Sorry line numbers are a little off above - the code had the
patches I sent you inside ;-)  Tell me if you want this reproduced in
plain ma9.

HTH Ralf

BTW:  Does your newsreader cut Subject: lines?

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