On Mon, Aug 31, 2009 at 4:30 PM, Ganesh Sittampalam<[email protected]> wrote: > These three patches between them fix the witnesses build on GHC 6.8.2, thus > resolving issue 1562. I'd have mentioned this in the patch message but I'm > not really sure which patch to put it in given that they are all relevant > (and I used different patches to get better commutation behaviour). > > The underlying issue is that the normal build specifies a whole bunch of > extensions in darcs.cabal, but the witnesses build doesn't. So a lot of > files rely on those extensions without specifying them explicitly in > LANGUAGE pragmas. Ideally we would specify everything in LANGUAGE pragmas, > but that might be significant effort for the really commonly used ones. > > On a side note, now that we're at GHC 6.8 as a minimum, can't we remove the > OPTIONS_GHC pragmas?
I'm okay with that. It will make it increasingly difficult for anyone who randomly is forced to build on ghc6.6 though. To me that is the only consideration and it's not worth the effort because we can tell them to use -fglasgow-exts. > Sat Aug 29 07:12:08 BST 2009 Ganesh Sittampalam <[email protected]> > * missing LANGUAGE in Darcs.Commands.Unrevert > > Sat Aug 29 07:12:56 BST 2009 Ganesh Sittampalam <[email protected]> > * missing LANGUAGE in Darcs.ProgressPatches > > Sat Aug 29 07:19:22 BST 2009 Ganesh Sittampalam <[email protected]> > * specify CPP globally for witnesses build These look great to me. I just can't push anything, yet. Trent could you push these please? It's my name on the review if they break something :) Thanks, Jason _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
