On Mon, 15 Oct 2012 15:13:58 +0200, Philipp Kern <pkern@debian.org> wrote: > On Mon, Oct 15, 2012 at 01:33:16AM +0200, Cyril Brulebois wrote: > > Stephen Kitt <steve@sk2.org> (14/10/2012): > > > Now that gcc-4.6 4.6.3-12 is installed in unstable on all > > > architectures, would it be possible to give gcc-mingw-w64 back on > > > all buildds? This will cause it to be rebuilt using gcc-4.6 > > > 4.6.3-12; since the latter's version ends up in the resulting binary > > > packages' versions, a binNMU shouldn't be necessary... > > > > > > gb gcc-mingw-w64_7 . ALL > > a package which failed to build can be given back. That's really an > > alias for “please give it another chance to build (successfully)”. If > > you want to get a(n already successfully built) package rebuilt > > against a new set of packages, that's where binNMUs come into play. > > And if would've been cool if you could send a rationale for why the binNMU > is needed (I simply don't know why one has to recompile against a new gcc, > and there's no bug# reference) to debian-release@lists.d.o (Cc'ed). > > (binNMUs migrate automatically if present, hence it's also slightly relevant > for wheezy release management.) Thanks for all the info everyone! Had I known a binNMU was the appropriate solution I would of course have filed a bug using reportbug with a rationale... Basically Matthias Klose asked me to update gcc-mingw-w64 to the latest version of gcc-4.6 which fixes the kfreebsd-amd64 build bug (#637236). This would also give us all the fixes in gcc-4.6 since the last gcc-mingw-w64 build (4.6.3-8), updates which warranted a freeze exception for gcc-4.6 (4.6.3-11 at least). As Jonathan pointed out, there's an RC bug against gcc-mingw-w64 (#690148), which I wasn't aware of because it got sent to the previous maintainer. I'll fix that and do a sourceful upload. Regards, Stephen
Attachment:
signature.asc
Description: PGP signature