Hi, Am Freitag, den 18.03.2011, 11:49 +0100 schrieb Giovanni Mascellani: > On 18/03/2011 05:04, Joachim Breitner wrote: > > Also, should we move to unstable with a regular upload, where the ghc > > packages are built by the buildd’s against ghc6, or by somehow copying > > existing ghc-7 binaries to unstable and _then_ doing a sourceful upload? > > Previously I thought the latter might be easier, but I think our package > > actually builds find against ghc6, so I guess we can go with the former. > > How would these two hypotheses relate with the libgmp3-related issue? In > unstable we still have libgmp3-dev, but early or late libgmp10-dev will > enter sid and we'll have to do again the Provides trick. I'm not fully > aware of the implications this will have: other than asking GMP > maintainers to keep the Provides against libgmp3-dev (or adding it again > when we'll need to do the bootstrap with the new dependencies), I think > this will require a rebuild of all Haskell packages. Am I correct? > > Moreover, how do you practically implement your second alternative? And > wouldn't it cause problems because of the GMP version disparity? > > Personally, I find the first option more clean and I would prefer it, > but I'm not sure I'm not missing anything in my evaluation. we have libgmp10-dev in unstable already, libgmp3-dev is gone. But as libgmp10-dev in unstable has the Provides: trick, ghc6 is installable in sid, so the initial unstable upload of ghc-7 will work. Once that is built on all arches, we can tell the gmp maintainer that he may drop the Provides. Or maybe I am not fully getting your point. Greetings, Joachim -- Joachim "nomeata" Breitner Debian Developer nomeata@debian.org | ICQ# 74513189 | GPG-Keyid: 4743206C JID: nomeata@joachim-breitner.de | http://people.debian.org/~nomeata
Attachment:
signature.asc
Description: This is a digitally signed message part