1. 29 May, 2009 2 commits
    • simonpj@microsoft.com's avatar
      Implement -XMonoLocalBinds: a radical new flag · 903831d5
      simonpj@microsoft.com authored
      The new flag -XMonoLocalBinds tells GHC not to generalise nested
      bindings in let or where clauses, unless there is a type signature,
      in which case we use it.  
      I'm thinking about whether this might actually be a good direction for
      Haskell go to in, although it seems pretty radical.  Anyway, the flag
      is easy to implement (look at how few lines change), and having it
      will allow us to experiement with and without.
      Just for the record, below are the changes required in the boot 
      libraries -- ie the places where.  Not quite as minimal as I'd hoped,
      but the changes fall into a few standard patterns, and most represent
      (in my opinion) sytlistic improvements.  I will not push these patches,
      == running darcs what -s --repodir libraries/base
      M ./Control/Arrow.hs -2 +4
      M ./Data/Data.hs -7 +22
      M ./System/IO/Error.hs +1
      M ./Text/ParserCombinators/ReadP.hs +1
      == running darcs what -s --repodir libraries/bytestring
      M ./Data/ByteString/Char8.hs -1 +2
      M ./Data/ByteString/Unsafe.hs +1
      == running darcs what -s --repodir libraries/Cabal
      M ./Distribution/PackageDescription.hs -2 +6
      M ./Distribution/PackageDescription/Check.hs +3
      M ./Distribution/PackageDescription/Configuration.hs -1 +3
      M ./Distribution/ParseUtils.hs -2 +4
      M ./Distribution/Simple/Command.hs -1 +4
      M ./Distribution/Simple/Setup.hs -12 +24
      M ./Distribution/Simple/UserHooks.hs -1 +5
      == running darcs what -s --repodir libraries/containers
      M ./Data/IntMap.hs -2 +2
      == running darcs what -s --repodir libraries/dph
      M ./dph-base/Data/Array/Parallel/Arr/BBArr.hs -1 +3
      M ./dph-base/Data/Array/Parallel/Arr/BUArr.hs -2 +4
      M ./dph-prim-par/Data/Array/Parallel/Unlifted/Distributed/Arrays.hs -6 +10
      M ./dph-prim-par/Data/Array/Parallel/Unlifted/Distributed/Combinators.hs -3 +6
      M ./dph-prim-seq/Data/Array/Parallel/Unlifted/Sequential/Flat/Permute.hs -2 +4
      == running darcs what -s --repodir libraries/syb
      M ./Data/Generics/Twins.hs -5 +18
    • simonpj@microsoft.com's avatar
      Fix Trac #3259: expose 'lazy' only after generating interface files · 0abcc755
      simonpj@microsoft.com authored
      This patch fixes an insidious and long-standing bug in the way that
      parallelism is handled in GHC.  See Note [lazyId magic] in MkId.
      Here's the diagnosis, copied from the Trac ticket.  par is defined 
      in GHC.Conc thus:
          {-# INLINE par  #-}
          par :: a -> b -> b
          par  x y = case (par# x) of { _ -> lazy y }
          -- The reason for the strange "lazy" call is that it fools the
          -- compiler into thinking that pseq and par are non-strict in
          -- their second argument (even if it inlines pseq/par at the call
          -- site).  If it thinks par is strict in "y", then it often
          -- evaluates "y" before "x", which is totally wrong.
      The function lazy is the identity function, but it is inlined only
      after strictness analysis, and (via some magic) pretends to be
      lazy. Hence par pretends to be lazy too.
      The trouble is that both par and lazy are inlined into your definition
      of parallelise, so that the unfolding for parallelise (exposed in
      Parallelise.hi) does not use lazy at all. Then when compiling Main,
      parallelise is in turn inlined (before strictness analysis), and so
      the strictness analyser sees too much.
      This was all sloppy thinking on my part. Inlining lazy after
      strictness analysis works fine for the current module, but not for
      importing modules.
      The fix implemented by this patch is to inline 'lazy' in CorePrep,
      not in WorkWrap. That way interface files never see the inlined version.
      The downside is that a little less optimisation may happen on programs
      that use 'lazy'.  And you'll only see this in the results -ddump-prep
      not in -ddump-simpl.  So KEEP AN EYE OUT (Simon and Satnam especially).
      Still, it should work properly now.  Certainly fixes #3259.
  2. 28 May, 2009 15 commits
  3. 27 May, 2009 1 commit
  4. 26 May, 2009 1 commit
  5. 19 May, 2009 2 commits
  6. 18 May, 2009 1 commit
    • Ben.Lippmeier@anu.edu.au's avatar
      Split Reg into vreg/hreg and add register pairs · f9288086
      Ben.Lippmeier@anu.edu.au authored
       * The old Reg type is now split into VirtualReg and RealReg.
       * For the graph coloring allocator, the type of the register graph
         is now (Graph VirtualReg RegClass RealReg), which shows that it colors
         in nodes representing virtual regs with colors representing real regs.
         (as was intended)  
       * RealReg contains two contructors, RealRegSingle and RealRegPair,
         where RealRegPair is used to represent a SPARC double reg 
         constructed from two single precision FP regs. 
       * On SPARC we can now allocate double regs into an arbitrary register
         pair, instead of reserving some reg ranges to only hold float/double values. 
  7. 21 Apr, 2009 1 commit
  8. 27 May, 2009 5 commits
    • simonpj@microsoft.com's avatar
      Fix Trac #3221: renamer warnings for deriving clauses · 685f6314
      simonpj@microsoft.com authored
      This patch arranges to gather the variables used by 'deriving' clauses,
      so that unused bindings are correctly reported.
    • simonpj@microsoft.com's avatar
      Template Haskell: allow type splices · 389cca21
      simonpj@microsoft.com authored
      At last!  Trac #1476 and #3177
      This patch extends Template Haskell by allowing splices in
      types.  For example
        f :: Int -> $(burble 3)
      A type splice should work anywhere a type is expected.  This feature
      has been long requested, and quite a while ago I'd re-engineered the
      type checker to make it easier, but had never got around to finishing
      the job.  With luck, this does it.
      There's a ToDo in the HsSpliceTy case of RnTypes.rnHsType, where I
      am not dealing properly with the used variables; but that's awaiting
      the refactoring of the way we report unused names.
    • simonpj@microsoft.com's avatar
      Template Haskell: improve lifting for strings · 97a8fe87
      simonpj@microsoft.com authored
      When you have a (\s::String -> ....[| s |]....), the string 
      's' is lifted.  We used to get a chain of single-character 
      Cons nodes, correct but lots and lots of code.  
      This patch arranges to optimise that to a string literal. It does
      so in two places:
        a) In TcExpr, if we know that s::String, we generate liftString directly
        b) In DsMeta, if we find a list of character literals, we convert to
           a string.  This catches a few cases that (a) does not
      There an accompanying  patch in the template-haskell package, 
      adding Language.Haskell.TH.Syntax.liftString
    • simonpj@microsoft.com's avatar
    • simonpj@microsoft.com's avatar
      Comments about wiredInIds · e598b8f9
      simonpj@microsoft.com authored
  9. 26 May, 2009 1 commit
  10. 24 May, 2009 2 commits
  11. 20 May, 2009 2 commits
  12. 19 May, 2009 2 commits
    • Duncan Coutts's avatar
      When linking a shared library with --make, always do the link step · 8aaa9ef4
      Duncan Coutts authored
      Without -shared, the default target is a binary and in that case
      it makes sense for --make to not try and link a binary when there
      is no Main module. But for a shared library the user already has
      to specify -shared and there's no reason a shared lib should
      contain any Main module or main function.
    • Duncan Coutts's avatar
      Make -dynload sysdep mean to embed rpaths in shared libs as well as binaries · 527f52a7
      Duncan Coutts authored
      Previously it only did it for binaries. This was presumably on the
      theory that the binary could specify the rpath for all the libs.
      However when it is the shared lib that is the final product (ie to
      link into a bigger project) then we need the rpaths for the shared
      lib to be self-contianed.
  13. 17 May, 2009 1 commit
  14. 15 May, 2009 2 commits
    • Duncan Coutts's avatar
      Set the soname when creating a shared lib · 6efacfe8
      Duncan Coutts authored
      It's still possible to override it, just use -optl-Wl,-soname, eg:
      ghc -shared -dynamic foo.o -o libfoo.so -optl-Wl,-soname,libbar.so
    • Duncan Coutts's avatar
      Keep C main separate from rts lib and link it in for standalone progs · fa00cc50
      Duncan Coutts authored
      Previously the object code for the C main function lived in the rts
      lib, however this is a problem when the rts is built as a shared lib.
      With Windows DLLs it always causes problems while on ELF systems it's a
      problem when the user decides to use their own C main function rather
      than a Haskell Main.main. So instead we now put main in it's own tiny
      little static lib libHSrtsmain.a which we install next to the rts libs.
      Whenever ghc links a program (without -no-hs-main) then it also links
      in -lHSrtsmain. For consistency we always do it this way now rather
      than trying to do it differently for static vs shared libraries.
  15. 14 May, 2009 1 commit
    • Duncan Coutts's avatar
      Remove old Windows-only implementation of keeping main outside the rts · 3d411991
      Duncan Coutts authored
      We now do it for all ways and for all platforms. This was a Windows-only
      version that only kept a separate Main.dyn_o for the dynamic linking case.
      It had to do that because Windows DLLs are stricter about unresolved symbols
      where as for ELF platforms we only run into the problem when we're not using
      a Haskell main function.
  16. 20 May, 2009 1 commit