On Sat, Nov 27, 2004 at 10:20:33AM +0100, Yann Dirson wrote: > > > I'm quite annoyed that this now becomes an RC issue for bigloo itself. > > > Especially since the bigloo build issue on m68k is preventing a new skribe > > > to enter sarge (the mips issue is due to bigloo being yet again autobuilt on > > > "reconfig", which is known to fail most of the time, and the mipsel failure > > > is an obvious chroot breakage). > > > > > Release people, would it be possible to get an exemption for the support of > > > bigloo on m68k for sarge ? Only skribe build-depends on it currently (and > > > it is quite unlikely IMHO that anyone uses bigloo on this arch, but I > > > suspect this one argument would be disregarded ;). > > You would need to file a bug on ftp.debian.org for this, and get an > > ftp-master to give you such an exemption. But unless the cause of the m68k > > build failure is known, I think an exemption is unlikely. > OK > > It looks to me > > like re-trying the build on a different buildd might help on this arch as > > well? > Since the killed process gets so because of an oom condition, we already > tried to rebuild on the buildd with the most ram, with no result. > And since the build fails with a different error (segfault) when using libgc > from the debian package, I suspect a gc bug. If this diagnostic is correct, > sure it would be better to fix the gc... however, this investigation needs > time, and it would not be fair to our users to avoid shipping the package > just because of this m68k-specific issue. All we'd manage to do is getting > this little-used arch more hate from users of other archs, which would not > be a good hing either... Build failures caused by bugs in dependent packages are one of the better reasons to ask for removal of stale binaries, IMHO. Please be sure to include this information in your bug report on ftp.debian.org. Thanks, -- Steve Langasek postmodern programmer
Attachment:
signature.asc
Description: Digital signature