On 26/10/10 at 10:55 +0100, Simon McVittie wrote: > tags 593023 + unreproducible > thanks > > It seems we have: > > * Lucas had one failed build in unstable > * Salvatore had one failed and one successful build in testing, and some > successful builds in unstable > * Serafeim had a successful build in unstable > > so it's probably going to be difficult to reproduce this. > > Youness (the upstream maintainer) reckons he isn't going to be able to debug > this without a backtrace and probably a tcpdump, so we're unlikely to make > progress on this unless someone can reproduce it. > > Lucas, would it be possible for you to rebuild libnice a few times > in your buildbot, preferably with tcpdump capturing traffic to and from > localhost? We can probably modify libnice to log a backtrace when the tests > fail, which would cover the other requirement.
It seems that the failure is timing-dependent. I run several builds concurrently on each machine, which causes quite some load. In my latest rebuild, it succeeded to build. I think that the severity could be lowered to !RC since we know that the package is buildable at least on some conditions. > Also, how unusual is your buildbot environment, not much. > and can you provide > instructions to reproduce the package set (presumably "use sbuild"?) yes, sbuild > , the > firewall it uses to deny Internet access for the build, and so on? simple local iptables rules. - Lucas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
