Re: Hackage-server: GHC 7

Just to follow-up on this - Most things are pretty easy to get this working, all of the code tweaks needed I believe have been pushed out. The rest of the changes will be to the package description, which I'm holding off on for now. The big holdups are getting happstack running. Happstack HEAD with syb-with-class HEAD both work on GHC 7, but neither have been released. I also have a patch against happstack-stable: http://groups.google.com/group/happs/browse_thread/thread/ad8497bec62f6b96# But I don't know if there will be another release from the 0.5 series. There should be a problem running any of the changes on GCH 6.12, and 6.10 compatibility wouldn't be hard (as long as happstack keeps the CPP hacks in place). Antoine On Sun, Nov 7, 2010 at 10:53 PM, Matthew Gruen wrote:

On Wed, Nov 3, 2010 at 6:08 AM, Duncan Coutts wrote:

...

On Sat, 2010-09-25 at 22:01 -0500, Antoine Latter wrote:

...

Hi folks,

I've been working on seeing what it takes to get hackage-server running on GHC 7 + the new libs that ship with it. For the most part, it's been pretty painless - a lot easier than the move to 6.12 was last year. It helps that the biggest breaking change (no longer generalizing non top-level bindings) is fixed by adding the ocasional type signature, which won't break any older versions.

Happstack requires a few fixes, though, which I'm requesting be pushed against happstack-head.  Would anyone be inconvenienced hackage-server required happstack-head? They plan on cutting a release around October.

Sounds ok to me. Ok Matthew?

This sounds like a great idea :) I may have a spot of quiet for the next few weeks (between midterm season and final project panicking), and I'm taking some time to see what kinks can be ironed out. Switching to head will hopefully do some ironing for me, though I think it's mostly SimpleHTTP that's improved, not so much state.

/Matt