1. 09 Jun, 2003 6 commits
  2. 08 Jun, 2003 2 commits
  3. 07 Jun, 2003 1 commit
    • ross's avatar
      [project @ 2003-06-07 11:20:47 by ross] · e8b2fe55
      ross authored
      Add $($(HcFlavour)_HAPPY_OPTS) to HAPPY_OPTS, where HcFlavour defaults to GHC.
      Packages may GHC_HAPPY_OPTS if they don't want -agc with ghc.
  4. 06 Jun, 2003 7 commits
  5. 05 Jun, 2003 7 commits
    • reid's avatar
      [project @ 2003-06-05 16:50:30 by reid] · c01ea09a
      reid authored
      Swap rules for building .o files by compiling and by assembling so that we compile if possible
    • reid's avatar
      [project @ 2003-06-05 14:39:44 by reid] · 78b1473d
      reid authored
      Put GC_C_OBJS into OBJS not STUBOBJS
    • reid's avatar
      [project @ 2003-06-05 14:05:36 by reid] · 7a69908a
      reid authored
    • stolz's avatar
      [project @ 2003-06-05 14:04:28 by stolz] · 7219d842
      stolz authored
      Work around broken getProcessStatus in 5.04.x
    • reid's avatar
      [project @ 2003-06-05 12:29:31 by reid] · dfc15b64
      reid authored
    • panne's avatar
      [project @ 2003-06-05 12:09:52 by panne] · 980c9615
      panne authored
      Attention: Tentative fix only! But at least this works much better
      than before.
      From the Linux man page for waitpid:
                ECHILD if  the  process specified in pid does not exist or
                       is not a child of the calling process.   (This  can
                       happen  for  one's  own  child  if  the  action for
                       SIGCHLD is set to SIG_IGN. See also the LINUX NOTES
                       section about threads.)
      Consequently, we don't ignore SIGCHLD anymore. Further changes:
         * SIGCONT is not ignored anymore, either. What was this for?
         * Don't use WUNTRACED in waitpid. Again, what was this for?
    • simonmar's avatar
      [project @ 2003-06-05 10:11:22 by simonmar] · 2ce1b8d7
      simonmar authored
      - Fix a couple of bugs in yesterday's pipeline cleanup.
      - Do some more tidying: share the code for filename generation between
        the two entry points to HscMain, and also share some of the other
        machinery in the Hsc phase.  This fixes some wibbles (things that
        were done in --make mode but not in one-shot, and vice-versa).
        One thing that works now is that if you say 'ghc -keep-hc-files Foo.hs',
        then it automatically switches to -fvia-C mode to generate the .hc file.
  6. 04 Jun, 2003 5 commits
    • stolz's avatar
      [project @ 2003-06-04 16:52:06 by stolz] · 3a47678a
      stolz authored
      System.Cmd.system-less invocation of sub-tasks
      (Sven approved reverting the OpenGL-Makefile change)
    • simonmar's avatar
      [project @ 2003-06-04 15:47:58 by simonmar] · 70b6c54b
      simonmar authored
      Grrr, started off making a small bugfix and ended up doing a major
      cleanup operartion.
      Anyway, the problem was that -odir wasn't putting the object files in
      the right place when the module in question has a hierarchical name.
      This was due to the object filename being generated in two different
      places: once by the compilation pipeline machinery, and again in the
      Finder.  It now works properly when --make is used; I haven't managed
      to fix it for one-shot compilations though (some replumbing is
      While I was here, I cleaned up the compilation pipeline machinery
      somewhat.  The previous scheme of generating a data structure
      representing the phases that need to be executed before actually
      executing them was wrong because the structure of the pipeline can
      change while it is being executed (eg. if we see {-# OPTIONS -fasm #-}
      during the CPP phase).  There were various hacks to deal with this,
      but it turned out to be quite messy.
      So the new story is that each compilation phase returns the name of
      the next phase to execute, and also figures out which file to put its
      output in.  This unfortunately means that the knowledge about what
      phases are done in what order is now spread throughout the module, but
      there are fewer hacks at the higher levels, and overall it seems to be
      an improvement.
    • panne's avatar
      [project @ 2003-06-04 15:18:29 by panne] · 4d65dec8
      panne authored
      Tentative fixes for Cywgin/MinGW-style pathname confusion
    • panne's avatar
      [project @ 2003-06-04 15:11:17 by panne] · ac1c3b11
      panne authored
      Always check for X Window System, otherwise $no_x is not set to "yes"
      when there is no X11. Confusing...
    • reid's avatar
      [project @ 2003-06-04 12:37:09 by reid] · 3a41699a
      reid authored
      All changes in this commit are standalone changes only but could usefully be used inplace as well: Don't store package.conf.installed; use a pipe to avoid problems from using stale package.confs.  Also, use PACKAGE, PACKAGE_DEPS, LIBRARY and LIBDIR settings from Makefile when installing pkg.  This requires some ugly sed hackery to overcome ghc-pkg not allowing trailing commas in a list: [a,b,c,]
  7. 03 Jun, 2003 4 commits
  8. 02 Jun, 2003 8 commits