I've committed what I believe is a fix for this, upstream; cf.:
bzr log -r revid:roz...@geekspace.com-20110313213224-442d4nti6k5el1lm
bzr diff -c revid:roz...@geekspace.com-20110313213224-442d4nti6k5el1lm
(there are a few other (minor) issues also being addressed upstream before
we have another release)
More background on this issue:
I actually couldn't reproduce this myself (I tried on Lenny, Squeeze,
and Sid (all amd64)--using both my own builds and the Debian builds)...,
HOWEVER, the presence of Curl_resolv_timeout at the outermost layer
in the stack-trace stood out; so I dug around for information on that,
and found things like:
http://curl.haxx.se/mail/tracker-2010-10/0054.html
... which led me to things like:
http://curl.haxx.se/mail/lib-2002-12/0103.html
So, we're apparently getting clobbered by alarm()+multithreading.
If I'd been able to figure out how to get the resolver to timeout
in the same way it was timing-out for Timo, I imagine I would have
seen the same crash.
--
"Don't be afraid to ask (λf.((λx.xx) (λr.f(rr))))."
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
