* On Friday, December 17 2010, Brent Yorgey wrote:
OK, I am confused. I grabbed the darcs repo corresponding to the '0.9.1' tag (just by doing a local 'get' from my up-to-date darcs repo). Then pulling in the single patch I listed above was easy enough (actually it pulled in two patches but the other patch was similar). Then I tried building it (against the released xmonad-0.9.1), which failed on some module because 'ExtensionClass' was not in scope.
After a bit more digging, I found that XMonad.Util.ExtensibleState is NOT in xmonad-contrib-0.9.1 as released on Hackage, but it IS in the darcs repo corresponding to the '0.9.1' tag. So now it seems that the 0.9.1 tag does not correspond to the actual 0.9.1 release, or else I am very confused (or both). So I am not sure what to do. Any thoughts?
Also confusing is the fact that Hackage reports xmonad-contrib-0.9.1 builds successfully with ghc 6.12, but as far as I know it shouldn't (unless the Hackage build process chooses dependency versions in a different way than cabal install, which is possible).
-Brent
I'm not sure if that's sjanssen's use of 'darcs tag', or whether we've misunderstood "darcs put --tag=0\.9\.1". I vaguely recall something similar happening the last time we had a release. The hackage xmonad-contrib 0.9.1 definitely works with ghc612. We could just cabal upload versions including the nice patches made by this distro, whose maintainers have already gone to the trouble of picking out right the changes: http://repos.archlinux.org/wsvn/community/xmonad-contrib/trunk/ghc-7-compat.... http://repos.archlinux.org/wsvn/community/xmonad/trunk/ghc-6.12-warnings.pat... -- Adam