#8139: ghc 7.6.3 fails to build on OS X 10.9 -------------------------------------+------------------------------------- Reporter: darinmorrison | Owner: Type: bug | Status: new Priority: normal | Milestone: 7.8.1 Component: Compiler | Version: 7.7 Resolution: | Keywords: Operating System: MacOS X | Architecture: x86_64 (amd64) Type of failure: Building GHC | Difficulty: Easy (less than 1 failed | hour) Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------- Changes (by darinmorrison): * status: closed => new * failure: None/Unknown => Building GHC failed * resolution: wontfix => * difficulty: Unknown => Easy (less than 1 hour) * version: 7.6.3 => 7.7 * milestone: => 7.8.1 * os: Unknown/Multiple => MacOS X Comment: I've reopening this and changed to 7.7; you seem to have a misunderstanding about the issues reported.
Why can't you use the binary build of ghc 7.6.3, and patch the settings file to point to a clang wrapper, as per a variant of Luke Ianni's hack? https://gist.github.com/lukexi/e11668ced04524564a8b . Based on my understanding, that should work with the newest xcode dev preview.
Your hack would only be a stopgap: the issues described prevent building of HEAD on 10.9 as well. Unless these patches, or some form of them, are commited, other developers who try to compile current and future version of ghc on 10.9 are going to run into the same problems. This is regardless of whether building with clang or gcc. thoughtpolice's clang preprocessor fixes (http://ghc.haskell.org/trac/ghc/ticket/7678) don't fix them. That means: * the mkdirhier.sh bug is still there on HEAD when compiling on 10.9 * the RtsProbe.d compilation bug is still there HEAD when compiling on 10.9 Those need to be fixed.
NB: apple doesn't support GCC anymore, as of xcode 5.
Accordingly, unless some core GHC devs have the time to do a 7.6.4 release (which they dont, and this has enough work arounds that I think it
NB: I've pointed out several times now that these issues happen with clang and with HEAD. See lukexi's notes compiling HEAD with clang on 10.9: https://gist.github.com/lukexi/6191728 probably cant be considered show stopping), i'm closing this as a wontfix. I will continue my discussion in #ghc with other developers. If need be, I can submit a more direct or modified patch. It's unacceptable to simply ignore these issues when they break compilation of both current and future versions of ghc on 10.9. They require a proper fix. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8139#comment:3 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler