Skip to content
Snippets Groups Projects
  1. Mar 10, 2014
  2. Feb 28, 2014
    • Austin Seipp's avatar
      binary-dist: when using xz, use extreme compression. · 251b18ab
      Austin Seipp authored
      
      When building a binary distribution with TAR_COMP=xz, using the -9e flag
      (extremely high compression) results in substantial savings: for the
      Mavericks builds, bzip2 scores in at about 120mb, while xz at level 9
      scores about 60mb - a huge reduction!
      
      This of course takes significantly longer - but it does not affect
      decompression speed for end users, so it's certainly worth it.
      
      Signed-off-by: default avatarAustin Seipp <austin@well-typed.com>
      251b18ab
  3. Feb 26, 2014
  4. Feb 17, 2014
  5. Jan 30, 2014
  6. Jan 28, 2014
  7. Jan 17, 2014
  8. Jan 14, 2014
  9. Jan 13, 2014
  10. Jan 07, 2014
  11. Nov 22, 2013
    • Austin Seipp's avatar
      Temporarily lower unfolding threshold on Windows · 9bb909bb
      Austin Seipp authored
      
      This is a very temporary, very unsatisfactory hack to fix #5987
      (for now.)
      
      The included comments essentially say it all: we lower the unfolding
      threshold to minimize some amount of exported symbols from the GHC
      stage2 DLL. I unfortunately had to lower it quite substantially for the
      dynamic stage2 build to pass.
      
      As of this writing, the DLL split between ghc.dll and ghc-0.dll is
      something like 26,000 vs 63,000 exported symbols, respectively. So we're
      still quite in danger of tripping it, but I think we will be OK at this
      exact moment.
      
      Signed-off-by: default avatarAustin Seipp <austin@well-typed.com>
      9bb909bb
  12. Oct 25, 2013
  13. Oct 01, 2013
  14. Sep 12, 2013
    • Austin Seipp's avatar
      Implement the AMP warning (#8004) · 75a9664a
      Austin Seipp authored
      
      This patch implements a warning when definitions conflict with the
      Applicative-Monad Proposal (AMP), described in #8004. Namely, this will
      cause a warning iff:
      
          * You have an instance of Monad, but not Applicative
          * You have an instance of MonadPlus, but not Alternative
          * You locally defined a function named join, <*>, or pure.
      
      In GHC 7.10, these warnings will actually be enforced with superclass
      constraints through changes in base, so programs will fail to compile
      then.
      
      This warning is enabled by default. Unfortunately, not all of
      our upstream libraries have accepted the appropriate patches. So we
      temporarily fix ./validate by ignoring the AMP warning.
      
      Dan Rosén made an initial implementation of this change, and the
      remaining work was finished off by David Luposchainsky. I finally made
      some minor refactorings.
      
      Authored-by: default avatarDan Rosén <danr@chalmers.se>
      Authored-by: default avatarDavid Luposchainsky <dluposchainsky@gmail.com>
      Signed-off-by: default avatarAustin Seipp <austin@well-typed.com>
      75a9664a
  15. Sep 04, 2013
    • Austin Seipp's avatar
      Make sure -fcmm-sink is passed to Parser properly · 9e133b9d
      Austin Seipp authored
      
      Parser.hs needs to be compiled with -fcmm-sink on x86 platforms, so the
      register allocator doesn't run out of stack slots. Previously, we had to
      do some CPP hacks in order to emit an #ifdef into the file - this is
      because we preprocess it once up front, and run the preprocessor again
      when we compile it.
      
      There's two cases: the boostrap compiler is > 7.8, and the stage1 parser
      needs the flag, or the stage1 compiler is compiling the stage2
      Parser.hs, and needs the flag..
      
      The previous approach was super fragile with Clang. The more principled
      fix is to instead do this through the build system.
      
      This fixes #8182.
      
      Signed-off-by: default avatarAustin Seipp <aseipp@pobox.com>
      9e133b9d
    • Herbert Valerio Riedel's avatar
      Hack-fix build breakage on Linux/GCC from fc4856f9 · 5a3918fe
      Herbert Valerio Riedel authored
      For some reason, the new `GccIsClang` variable introduced via
      fc4856f9 for addressing #8148
      isn't set explicitly to `NO`; so this simply changes the test
      `ifeq $(GccIsClang) NO` to `ifneq $(GccIsClang) YES` which
      should fix the build with and w/o Clang.
      5a3918fe
    • chak@cse.unsw.edu.au.'s avatar
      Make validate play nice with clang (for Xcode 5 command line tools) · fc4856f9
      chak@cse.unsw.edu.au. authored
      * This partially fixes #8148. However, --with-ghc-4.8 will still not work given the rather dubious m4 macros and the failures in the test suite due to '-nodefaultlibs' still need to be fixed.
      fc4856f9
  16. Sep 03, 2013
  17. Aug 24, 2013
  18. Aug 14, 2013
  19. Aug 05, 2013
  20. Jul 03, 2013
    • Ian Lynagh's avatar
      Change the ranlib detection · c548fec4
      Ian Lynagh authored
      On Windows, the ranlib in the path may not be the right ranlib (it may
      be the 32bit ranlib when we are making a Win64 compiler, or vice-versa).
      Therefore we can't leave it up to libffi to detect the right ranlib, but
      need to tell it which ranlib to use. This means that we need to find
      ranlib even if we don't actually need it ourselves.
      c548fec4
  21. Jul 02, 2013
  22. Jun 19, 2013
  23. Jun 17, 2013
  24. Jun 14, 2013
  25. Jun 09, 2013
  26. May 31, 2013
  27. May 30, 2013
  28. May 04, 2013
  29. Apr 27, 2013
  30. Apr 21, 2013
  31. Apr 20, 2013
  32. Apr 03, 2013
Loading