Re: Feasibility of backports?
On Thu, Dec 8, 2011 at 1:15 PM, Joachim Breitner <firstname.lastname@example.org> wrote:
> Am Donnerstag, den 08.12.2011, 20:51 +0100 schrieb Iustin Pop:
>> I'm asking more from the point of view of upstream, rather than Debian
>> packaging. I presume that due to the ghc6→ghc migration, doing backports
>> for a few simpler packages (not yesod or such) is still not an easy
>> task, right?
>> A good example that I'm thinking about is aeson; it has about 5-6
>> dependencies (I have no idea if these have in turn more dependencies
>> which are not in squeeze), so I think it would take some effort but
>> would be doable.
> my thought is that if we do backports, then we should backport the
> complete set of haskell packages, including ghc, so the ghc6→ghc
> migration should not be a problem; we just do it in backports as well.
> So it is basically a problem of rebuilding everything, i.e. of
> Maybe, first someone should script something to rebuild ghc_7-* and
> haskell-* on a Debian stable machine and provide an unofficial backport.
> If that works out well and user demand is present, then we can consider
> an official backport.
Towards this end, I have used our autobuilder to create a squeeze backport at
deb http://deb.seereason.com/debian squeeze-seereason main
deb-src http://deb.seereason.com/debian squeeze-seereason main
There will be about 300 packages there when the upload finishes in a
few minutes. Most of the packaging is produced using a tool named
cabal-debian, so there are a few small packaging differences from the
standard packages in sid.