This a headsup that there is a ghc refresh update for EPEL 5 now in testing:
This updates ghc from 6.12.3 to a bit more recent stable release,
which is also the currently the version in EPEL6.
Later I am thinking to refresh EPEL6 to ghc-7.4.2.
(Current F20 has 7.6.3.)
If we had scl for EPEL it would be good to ship
multiple ghc versions over the EPEL lifetime,
since building newer packages with the older
ghc versions is often problematic.
Eg the latest pandoc releases needs ghc-7.4 to build.
But some people might still want to stay with
the older version for their projects.
Do test if you are interested or using EPEL5 ghc.
Happy New Year,
It has been quiet here for a while so I thought I would post
a few updates about what I have been doing recently.
* cabal-0.8.7 is out and in updates-testing now for 1 week.
Main new feature is an install command which tries to install
any missing system dependencies before running "cabal install".
(0.8.8 will make that a bit less strict so it doesn't fail
when there are dependencies which are not yet packaged.)
* ghc-7.0.4 and cabal-install for EPEL5
I spent a little at the end of last month backporting
ghc-7.0.4 from EPEL6 to EPEL5 (this may will be the last
ghc update I do for RHEL5 but we shall see).
Next I'd like to backport ghc-7.4.2 et al from F19 to EPEL6.
(Currently stable EPEL 5 and 6 have 6.12.3 and 7.0.4.)
* Haskell Platform and ghc update
2013.4 didn't get released but it seems there will be a release
coming with various version bumps so I will probably wait for
that before starting all the version bumping work for F21.
Still no sight of a ghc-7.8 release yet: it may well be after
that HP release which will be based on ghc-7.6.3. Depending
on the timing of the ghc release I am still considering to
ship ghc-7.8 in F21, which would be the first Fedora release
not following HP for a long time - though Debian and Ubuntu
have already done this for some time now I believe.
* Triaging of URM bugs
There are a lot of version updates bugs open in bugzilla.
I spent some time gradually triaging them and there are now
currently only 9 more left to triage (belonging to others)
(there were dozens when I started). This time I am using
the Whiteboard field to tag updates for f21. So if one of your
package update bugs has f21 in the Whiteboard it means I believe
it is good to go into F21, but you don't need to build it yet
unless you are able and willing to rebuild/bump its reverse
dependencies too as usual. :)
So there are now 3 piles of updates bugs:
9 untriaged (empty Whiteboard)
57 triaged to f21
19 blocked updates
See https://fedoraproject.org/wiki/Haskell_SIG#Bugs for the links.
The blocked updates mostly need new dependencies to be reviewed and added to Fedora.
So if one of your package is blocked on new deps in the Whiteboard
please file review requests to get those moving.
I have filed a few, but more are needed - most of them are mine though. ;o)
* Package reviews have been moving pretty slowly.
I made a bit more progress on unpackaging haskell-platform
largely thanks to Christopher Meng who is quite active
reviewing recently but more help is needed.
has 11 Ready New reviews waiting. Please help if you can.
More reviews are needed to unlock more updates:
- hslua blocks pandoc-1.12
- haskell-platform: parsec, syb, regex-compat, async (more to come)
- highlighting-kate rename
are the main open ones from my pov.
I am trying to find the time to rewrite my rebuild shellscripts
in haskell-sig.git to something more robust in Haskell and
hopefully use that for this next major round of updating.
I will post again if/when I make progress on this.
I think that is basically everything. General plans and schedule
for F21 still seem to be undecided. I am still thinking to
propose a Haskell Spin for F21 if possible: perhaps I should
post to the spins list asking about that.