On Wed, Jun 15, 2011 at 12:45:22PM +0200, Thijs Kinkhorst wrote: > On Thu, June 9, 2011 23:03, Peter Samuelson wrote: > > > > [Thijs Kinkhorst] > >> The last two security updates for subversion, 1.6.12dfsg-6 and > >> 1.6.12dfsg-7, > >> have failed to build on kfreebsd-i386 and kfreebsd-amd64. Attached are > >> two > >> sample build logs. Can you investigate this? > > > > It is a problem we've seen on _some_ but not all kfreebsd buildd hosts > > for quite some time, and last we tried, Cyril could not reproduce it on > > his own system. One of the test suites starts up a server process on > > localhost, connects to it, does some tests, and eventually kills the > > server. Starting the server, or connecting to it, is what appears to > > be failing. It works on all our Linux platforms, and it _used_ to work > > on some of the kfreebsd buildds as well, though lately it seems to not > > work on any of them. > > > > Is it valid to create a TCP based server on 127.0.0.1, connect to it, > > then later kill it, in a build chroot? If not, I guess I can disable > > all those tests. I've just always assumed that was a valid thing to do > > in a package build. > > Thanks. > I'm not a buildd expert, so I'm CC'ing some of those that are. > > If it builds at least on some buildds (which ones?), it would help me in > the sort term if we could get those buildds to build 1.6.12dfsg-6.
I think it builds fine on non-smp buildds, which at the time of the Squeeze release were fano and finzi. They have been switched to SMP a few weeks ago. Maybe we should switch them back to non-SMP? Alternatively I can build them manually to solve this situation, though it will only hide the problem. Aurelien -- Aurelien Jarno GPG: 1024D/F1BCDB73 aurel...@aurel32.net http://www.aurel32.net -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
