[arch-haskell] [arch-general] Haskell Support Was: Xmonad version?

[Magnus Therning]

On Wed, Dec 21, 2011 at 06:00:02PM +0100, Fabio Riga wrote:
> 2011/12/20 Magnus Therning <magnus at therning.org>
> 
>> Den 20 dec 2011 04:58 skrev "Bernardo Barros" <bernardobarros2 at gmail.com>:
>>
>>> Arch is not a testing distro... And it's not even up to date with
>>> Haskell Platform. If you want to "go faster them haskell platform"
>>> you will need to work hard with packaging testing and this seems
>>> NOT the case with Arch at the moment...
>>
>> I don't think it's quite that easy.  AFAIU we'll soon have a choice
>> to make, either stick to HP with an old version of GHC, or move to
>> the latest stable GHC (7.4) and drop HP proper.
>>
>> Personally, I'm in favour of the latter.
>>
>> /M
>>
> *From Arch Linux About page:*
> Arch strives to stay bleeding edge, and typically offers the latest stable
> versions of most software. [...]
> Arch Linux uses a "rolling release" system which allows one-time
> installation and perpetual software upgrades.
> ... and many other interesting feautres that made me love this distribution.
> 
> Haskell Platform is not bleeding edge, it seems to follow a "old
> versions are more stable" approach, more in the way of Debian. This
> approach has is merits, but I prefer the "Arch way", so I vote for
> dropping HP.
> 
> Actually Ghc is not the problem. Most packages in hackage are
> already builded with ghc-7.2, but 7.2 is testing, so we should use
> 7.04. The problem are the other packages: for example every update
> that depends upon * text* failed because in HP (until some days ago)
> the version was too old.  In a couple of months this package will be
> too old again and will break other packages, unless we update
> haskell packages twice a year.

I think we all agree that GHC isn't the problem per se, it just
created a very awkward situation when upstream decided that 7.2 (which
based on the version number is stable) was a "tech preview".  Lots of
people just grabbed the latest release with a even version number and
started using it.  This is probably a good thing for upstream (lots of
testing), but it isn't very good for packagers.  It does mean that a
lot of packages have recent releases that haven't been tested on 7.0,
which does affect us.  This situation is temporary though and I hope
upstream realises what effects their decision had.

I was in favour of switching to 7.2 despite it being a "tech preview".

> IMHO, if cabal installs the latest available hackage, we sholud
> simply do the same.

As far as possible yes, there will always be dependencies that will
make us lag at times, but without HP we won't have any upstream for
libraries that are on a fixed release schedule.  That will increase
our chances of sticking close to the edge AFAICS.

Dropping HP might have a big impact on [haskell] though.  At the
moment [extra]/[community] offers a stable base to build on, this will
go away in the future.  Every upgrade to [extra]/[community] has the
potential to render [haskell] un-buildable.  It will be interesting to
see how well we can communicate to avoid that :-)
It's still a net-win in my book though!

/M

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


Perl is another example of filling a tiny, short-term need, and then
being a real problem in the longer term.
     -- 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/20111221/df228413/attachment.pgp>