[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: ghc 6.10 and provides

On Sun, Apr 05, 2009 at 05:05:18PM +0100, Magnus Therning wrote:
> I might be behind the curve here, but what's the reason behind the
> changes in "virtual provides" between ghc 6.8 and 6.10?

Because I broke that and never got around to fix it and test that it
works, too.

Though now I'm thinking that it's not generally a good idea to expect
an one to one mapping from cabal names to libghc6-*-dev names anyway,
but better to look at what files you are using and what package they
belong to.  We may need to have multiple versions of a library around,
quickcheck{1,2} as an example.  Though in that particular case I set
quickcheck1 (and not quickcheck2) to have a Provides:
libghc6-quickcheck-dev, since it's still the version people use.
There may not always be a clear preferred version like that.

Never mind that we could well have avoided that particular scenario by
still leaving QuickCheck version 2 unpackaged...  We'd end up with
software using multiple versions of libraries sooner or later, anway.

I'll remember to enable those libghc6-*-dev virtual packages for
6.10.2, even though they mostly look like so much noise to me.

Reply to: