Skip to content
Commits on Oct 1, 2012
  1. @simonpj
  2. @simonpj
  3. @simonpj
  4. @simonpj

    Fix an ASSERT

    simonpj committed Oct 1, 2012
  5. @simonpj

    Use System Names for evidence variables

    These variables all have uniformative names like "cobox" or "d".
    Make them System Names tells the pretty printer to show their
    uniques even without -dppr-debug.  That in turn makes the
    less-voluminous debug traces *much* more informative.
    simonpj committed Oct 1, 2012
  6. @simonpj
  7. @simonpj
Commits on Sep 29, 2012
  1. @simonpj

    Add missing case in TcUnify.matchExpectedFunKind

    This fixes Trac #7278
    simonpj committed Sep 29, 2012
Commits on Sep 28, 2012
  1. @simonpj
  2. @simonpj

    Refactor the handling of kind errors

    * Treat kind-equality constraints as *derived* equalities,
      with no evidence.  That is really what they are at the moment.
    
    * Get rid of EvKindCast and friends.
    
    * Postpone kind errors properly to the constraint solver
      (lots of small knock-on effects)
    
    I moved SwapFlag to BasicTypes as well
    simonpj committed Sep 28, 2012
  3. @simonpj

    Better ASSERT message

    simonpj committed Sep 28, 2012
  4. @simonpj

    Comments only

    simonpj committed Sep 28, 2012
  5. @simonpj
  6. @simonpj

    Comments only

    simonpj committed Sep 28, 2012
  7. @simonpj

    Wibble to debug print

    simonpj committed Sep 28, 2012
  8. @simonpj
Commits on Sep 27, 2012
  1. @igfoo
  2. @igfoo
  3. @igfoo

    Tweak the build system handling of shell wrappers

    Rather than having a separate
        foo_INSTALL_SHELL_WRAPPER
    variable, we just use
        foo_INSTALL && foo_SHELL_WRAPPER
    igfoo committed Sep 27, 2012
  4. @igfoo
Commits on Sep 26, 2012
  1. @igfoo
  2. @ezyang

    Partially fix #367 by adding HpLim checks to entry with -fno-omit-yie…

    …lds.
    
    The current fix is relatively dumb as far as where to add HpLim
    checks: it will always perform a check unless we know that we're
    returning from a closure or we are doing a non let-no-escape case
    analysis.  The performance impact on the nofib suite looks like this:
    
                Min          +5.7%     -0.0%     -6.5%     -6.4%    -50.0%
                Max          +6.3%     +5.8%     +5.0%     +5.5%     +0.8%
     Geometric Mean          +6.2%     +0.1%     +0.5%     +0.5%     -0.8%
    
    Overall, the executable bloat is the biggest problem, so we keep the old
    omit-yields optimization on by default. Remember that if you need an
    interruptibility guarantee, you need to recompile all of your libraries
    with -fno-omit-yields.
    
    A better fix would involve only inserting the yields necessary to break
    loops; this is left as future work.
    
    Signed-off-by: Edward Z. Yang <ezyang@mit.edu>
    ezyang committed Sep 17, 2012
  3. @igfoo

    Build old-time with stage2

    It's no longer needed in order to build the compiler itself.
    igfoo committed Sep 26, 2012
  4. @simonmar

    fix stage2 compilation

    simonmar committed Sep 26, 2012
Commits on Sep 25, 2012
  1. @simonmar

    Remove some old-codegen cruft

    simonmar committed Sep 25, 2012
  2. @simonmar

    When -split-objs is on, make one SRT per split, not one per module

    This is a hopefully temporary measure until the new SRT design is
    implemeented.
    simonmar committed Sep 25, 2012
  3. @simonmar

    fix panic message typo

    simonmar committed Sep 25, 2012
Commits on Sep 24, 2012
  1. @simonmar

    Misc tidyup

    simonmar committed Sep 24, 2012
  2. @simonmar
  3. @simonmar

    whitespace only

    simonmar committed Sep 21, 2012
  4. @simonmar
  5. @simonmar

    no functional changes

    simonmar committed Sep 21, 2012
  6. @simonmar

    Another overhaul of the recent_activity / idle GC handling (#5991)

    Improvements:
    
     - we now turn off the timer signal in the non-threaded RTS after
       idleGCDelay.  This should make the xmonad users on #5991 happy.
    
     - we now turn off the timer signal after idleGCDelay even if the
       idle GC is disabled with +RTS -I0.
    
     - we now do *not* turn off the timer when profiling.
    
     - more comments to explain the meaning of the various ACTIVITY_*
       values
    simonmar committed Sep 21, 2012
  7. @simonmar

    Revert "Disable the timer signal while blocked in select() (#5991)"

    This reverts commit dd24d6b.
    
    This attempt to fix the problem was misguided: the program might be
    stuck in a foreign call rather than awaitEvent(), and then the timer
    signal will never get disabled.  The only way to turn off the timer
    signal in this case is in the timer interrupt handler itself.
    simonmar committed Sep 21, 2012
Commits on Sep 23, 2012
  1. @igfoo
Something went wrong with that request. Please try again.