Hi, Am Samstag, den 24.12.2011, 08:28 +1100 schrieb Erik de Castro Lopo:

Joachim Breitner wrote:

...

ghc-stage2: panic! (the 'impossible' happened) (GHC version 7.4.0.20111219 for powerpc-unknown-linux): Cant do annotations without GHCi {libraries/vector/Data/Vector/Fusion/Stream/Monadic.hs:104:19-33} base:GHC.Exts.ForceSpecConstr{d rbL}

Thats related to the problem of GHCi being broken on powerpc.

Somewhat: vector uses annotations, annotation need GHCi, GHCi is disabled (because broken) on powerpc, vector is bundled by accident in GHC (and will be removed in the final release, unless I am mistaken), and the real vector package is patched in Debian to handle this (http://patch-tracker.debian.org/patch/series/view/haskell-vector/0.9-2/no_an...). My hope is that ghc will in the future ignore annotations on non-GHCi-arches (http://hackage.haskell.org/trac/ghc/ticket/4268). Once vector is unbundled, I expect the build to succeed on powerpc. I guess the other weird-arch builds would have failed at this point as well, if they had not failed due to the various other reasons earlier; I hope these get fixed in the final release. Maybe Debian should offer GHC to run buildbots (http://hackage.haskell.org/trac/ghc/wiki/Builder) on at least some of our weird arch machines, so that these problems will be noticed. earlier. But I’m not sure if the DSA team allows such a thing, as we’d be running code that is not under our control. Is someone in the Haskell team interested in finding that out? Greetings, Joachim -- Joachim "nomeata" Breitner Debian Developer nomeata@debian.org | ICQ# 74513189 | GPG-Keyid: 4743206C JID: nomeata@joachim-breitner.de | http://people.debian.org/~nomeata