1. 02 Apr, 2004 1 commit
  2. 31 Mar, 2004 1 commit
    • simonmar's avatar
      [project @ 2004-03-31 15:23:16 by simonmar] · 17d537ba
      simonmar authored
      ClosureInfo.might_be_a_function: this predicate wasn't taking into
      account abstract types, which might also hide functions underneath.
      
      Fixes broken compiler when compiled without -O.
      17d537ba
  3. 30 Mar, 2004 2 commits
  4. 27 Mar, 2004 1 commit
  5. 24 Mar, 2004 2 commits
  6. 23 Mar, 2004 5 commits
  7. 22 Mar, 2004 4 commits
  8. 21 Mar, 2004 1 commit
  9. 19 Mar, 2004 4 commits
    • panne's avatar
      [project @ 2004-03-19 23:20:20 by panne] · 9b5e0886
      panne authored
      Fixed  the JOB line in heap  profiles, it contained superfluous spaces
      and an evil line break.
      
      Merge to STABLE
      (This fix looks quite right, but again, I leave this to the Master of
      Releases (tm), because there might already be tools depending on the
      slightly wrong old format.)
      9b5e0886
    • panne's avatar
      [project @ 2004-03-19 23:17:06 by panne] · 06ebaa8c
      panne authored
      Restrict the setlocale() to LC_CTYPE (was LC_ALL), because printf and friends
      are LC_NUMERIC-sensitive. This fixes the generation of heap profiles for e.g.
      de_DE locales where a comma separates the fractional part of a number, not a
      dot. (There was a similar bug in Hugs regarding atof, BTW, which was fixed last
      December.)
      
      Merge to STABLE
      (I leave this to the Grand Marshal of Releases, Mr. SimonM, because I'm not 100%
      sure of all the consequences.)
      06ebaa8c
    • simonpj's avatar
      [project @ 2004-03-19 13:25:24 by simonpj] · 45deb4d6
      simonpj authored
      Add comments
      45deb4d6
    • simonpj's avatar
      [project @ 2004-03-19 13:24:52 by simonpj] · fac53ab4
      simonpj authored
      Fix naming bogon for Typeable
      fac53ab4
  10. 18 Mar, 2004 3 commits
  11. 17 Mar, 2004 5 commits
    • simonpj's avatar
      [project @ 2004-03-17 15:38:13 by simonpj] · ad90960b
      simonpj authored
      Wibble
      ad90960b
    • simonpj's avatar
      [project @ 2004-03-17 13:59:06 by simonpj] · af5a2151
      simonpj authored
      ------------------------
      	More newtype clearing up
      	------------------------
      
      * Change the representation of TyCons so that it accurately reflects
      	* data     (0 or more constrs)
      	* newtype  (1 constr)
      	* abstract (unknown)
        Replaces DataConDetails and AlgTyConFlavour with AlgTyConRhs
      
      * Add IfaceSyn.IfaceConDecls, a kind of stripped-down analogue
        of AlgTyConRhs
      
      * Move NewOrData from BasicTypes to HsDecl (it's now an HsSyn thing)
      
      * Arrange that Type.newTypeRep and splitRecNewType_maybe unwrap just
        one layer of new-type-ness, leaving the caller to recurse.
      
        This still leaves typeRep and repType in Type.lhs; these functions
        are still vaguely disturbing and probably should get some attention.
      
      Lots of knock-on changes.  Fixes bug in ds054.
      af5a2151
    • simonpj's avatar
      [project @ 2004-03-17 10:44:06 by simonpj] · a34e79f1
      simonpj authored
      More Windows building notes
      a34e79f1
    • simonpj's avatar
      [project @ 2004-03-17 10:12:31 by simonpj] · 2407332d
      simonpj authored
      Fix debug-printing for Insts
      2407332d
    • simonpj's avatar
      [project @ 2004-03-17 08:54:19 by simonpj] · a6001411
      simonpj authored
      Use sub-kind rather than kind-equality check in matchTys,
      just as we do in TcUnify.  This is someting I failed to
      do when shifting to the funky new kind representation.
      a6001411
  12. 16 Mar, 2004 5 commits
  13. 15 Mar, 2004 1 commit
  14. 13 Mar, 2004 1 commit
  15. 12 Mar, 2004 2 commits
  16. 11 Mar, 2004 2 commits
    • simonpj's avatar
      [project @ 2004-03-11 14:34:22 by simonpj] · 9ec3012e
      simonpj authored
      Fix a nasty and long-standing bug in the handling of functional dependencies.
      
      The story is told in comments with TcSimplify.tcSimplifyRestricted.  We were
      simpifying a group of constraints *twice*: once to establish the type vars to
      quantify over, and once "for real" but less brutally.  Unfortunately, the
      less-brutally part meant that we did less improvement, which in turn meant
      that an invariant (no captured constraints) was violated.  Consequential
      bizarre results.
      
      The test is typecheck/should_compile/tc177
      9ec3012e
    • simonpj's avatar
      [project @ 2004-03-11 14:31:56 by simonpj] · 1796a476
      simonpj authored
      Better printing for LHsBinds
      1796a476