[arch-haskell] Haskell's state in Arch & near future

[Magnus Therning]

On Sun, Nov 06, 2011 at 12:28:29PM -0800, Leif Warner wrote:
> On Sun, Nov 6, 2011 at 1:16 AM, Magnus Therning <magnus at therning.org> wrote:
> 
> > On Sat, Nov 05, 2011 at 08:22:54PM +0100, Peter Hercek wrote:
> > [...]
> > > I personally do not mind if we would move quicker than haskell
> > > platform. Though it means additional work so it is up to the
> > > maintainers.
> >
> > The main problem with actually getting the speed we'd like is that the
> > speed of upstream developers is a limiting factor.  With each package
> > that is added to ArchHaskell we add yet another possible speed bump.
> > Finding a solution is tricky, since the only option that doesn't
> > require a lot of resources is dropping packages.
> >
> 
> I've noticed, when trying to update some things, that there's been a
> few packages moving past Haskell Platform and requiring GHC 7.2.1.

That's not such a big problem per se.  The real issue is when some
packages don't move up to 7.2.1 but we want to.  In the past I've sent
out emails to upstream maintainers to encourage them to update.  So
far that has worked fairly well, but I don't think that will be the
case always.  Then we're faced with the option of either keeping
everything thing back, update the package ourselves, or drop the
package completely.  Neither option is very good.

> > > Dropping all haskell packages from aur is a good idea. Maybe we
> > > should leave there only one "dummy" package which would point
> > > haskell newbies to a description how to get additional haskell
> > > packages which a archlinux user would expect to be in aur.
> >
> > IIRC yaourt can install packages directly from hackage.  Is that
> > still the case?
> > If so I'll be more than happy to disown all AUR packages and close
> > down the arch-haskell account.
> 
> Yaourt doesn't do that, afaik. That was bauerbill I think. It's nice
> having the Haskell packages on AUR, but not if they're not kept
> up-to-date.  Many of the arch-haskell PKGBUILDs that are "outdated"
> up there will still build fine if the dependencies on specific
> versions are removed.

Yes, you are right, it was bauerbill that did that.

> If we're going to the trouble to maintain working PKGBUILDs for the
> binary repo, why not sync those up to AUR, and disown the rest?

That's absolutely an option, though if others do add packages to AUR,
that we later add to [haskell] then we run the very real risk of
having Haskell packages on AUR that don't build with each other.
That's not a terrible situation in my opinion though.

/M

-- 
Magnus Therning                      OpenPGP: 0xAB4DFBA4 
email: magnus at therning.org   jabber: magnus at therning.org
twitter: magthe               http://therning.org/magnus

Most software today is very much like an Egyptian pyramid with
millions of bricks piled on top of each other, with no structural
integrity, but just done by brute force and thousands of slaves.
     -- Alan Kay
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://www.haskell.org/pipermail/arch-haskell/attachments/20111107/09818157/attachment.pgp>