1. 17 Jun, 2002 3 commits
  2. 16 Jun, 2002 1 commit
    • panne's avatar
      [project @ 2002-06-16 16:10:29 by panne] · f179e855
      panne authored
      * Use the encoded module name for the C stub of "foreign import
        wrapper", so e.g. the '.' in hierarchical module names gets mangled.
      * Nuke unused parameter in mkFExportCBits
  3. 15 Jun, 2002 1 commit
  4. 14 Jun, 2002 5 commits
  5. 12 Jun, 2002 2 commits
    • wolfgang's avatar
      [project @ 2002-06-12 22:29:43 by wolfgang] · f55855f2
      wolfgang authored
      Initial support for GHCi on MacOS X.
      The code still has problems with global variables in compiled C code, but
      it _seems_ to work OK for GHC-compiled code. Many error checks are still
    • wolfgang's avatar
      [project @ 2002-06-12 22:04:25 by wolfgang] · 643a2f70
      wolfgang authored
      Added support for Frameworks on MacOS X.
      *) On MacOS X, two additional command-line options are supported:
      -framework <FRAMEWORK>    link with framework, gets passed on to ld
      -framework-path <PATH>    gets passed on to ld as "-F<PATH>". (-F is already taken for GHC).
      *) Two corresponding additional options for package.conf files:
      These options are allowed on any platform. They are ignored everywhere except on MacOS X.
      *) ghc-pkg no longer uses Read. Instead, it uses a Happy parser. ghc/utils/ghc-pkg/ParsePkgConfLite.y is basically a copy of ghc/compiler/main/ParsePkgConf.y. "Light" refers to the fact that it doesn't depend on other GHC modules and has less sophisticated error reporting.
      *) Frameworks will need some additional work for GHCi.
  6. 09 Jun, 2002 2 commits
  7. 07 Jun, 2002 1 commit
    • chak's avatar
      [project @ 2002-06-07 07:16:04 by chak] · 2205f0ce
      chak authored
      Fixed handling of infix operators in types:
      - Pretty printing didn't take nested infix operators into account
      - Explicit parenthesis were ignored in the fixity parser:
        * I added a constructor `HsParTy' to `HsType' (in the spirit of `HsPar' in
          `HsExpr'), which tracks the use of explicit parenthesis
        * Occurences of `HsParTy' in type-ish things that are not quite types (like
          context predicates) are removed in `ParseUtils'; all other occurences of
          `HsParTy' are removed during type checking (just as it works with `HsPar')
  8. 06 Jun, 2002 1 commit
  9. 05 Jun, 2002 3 commits
    • simonpj's avatar
      [project @ 2002-06-05 14:39:27 by simonpj] · b2f644fa
      simonpj authored
      	Add rebindable syntax for do-notation
      		(this time, on the HEAD)
      Make do-notation use rebindable syntax, so that -fno-implicit-prelude
      makes do-notation use whatever (>>=), (>>), return, fail are in scope,
      rather than the Prelude versions.
      On the way, combine HsDo and HsDoOut into one constructor in HsSyn,
      and tidy up type checking of HsDo.
    • simonpj's avatar
      [project @ 2002-06-05 14:08:23 by simonpj] · 2145e55a
      simonpj authored
      	Fix the (new) lexer, and make the derived read
      	and show code work according to the new H98 report
      The new lexer, based on Koen's cunning parser (Text.ParserCombinators.ReadP)
      wasn't quite right.  It's all very cool now.
      In particular:
      * The H98 "lex" function should return the exact string parsed, and it
      now does, aided by the new combinator ReadP.gather.
      * As a result the Text.Read.Lex Lexeme type is much simpler than before
          data Lexeme
            = Char   Char	-- Quotes removed,
            | String String	-- 	escapes interpreted
            | Punc   String 	-- Punctuation, eg "(", "::"
            | Ident  String	-- Haskell identifiers, e.g. foo, baz
            | Symbol String	-- Haskell symbols, e.g. >>, %
            | Int Integer
            | Rat Rational
            | EOF
           deriving (Eq,Show)
      * Multi-character punctuation, like "::" was getting lexed as a Symbol,
      but it should be a Punc.
      * Parsing numbers wasn't quite right.  "1..n" got it confused because it
      got committed to a decimal point and then found a second '.'.
      * The new H98 spec for Show is there, which ignores associativity.
    • simonpj's avatar
      [project @ 2002-06-05 14:07:47 by simonpj] · 84e376bc
      simonpj authored
      Add comment
  10. 04 Jun, 2002 3 commits
  11. 03 Jun, 2002 2 commits
    • sof's avatar
      [project @ 2002-06-03 23:36:40 by sof] · a2fb599c
      sof authored
      whether or not to quote -D and -U opts is a property of the host, not target OS
    • matthewc's avatar
      [project @ 2002-06-03 13:08:37 by matthewc] · cb5ccf0a
      matthewc authored
      Initial mangling and tailcalls support for IA64.
      Function prologues and epilogues are deleted and we use a single register
      stack frame throughout (with a little register renaming in the mangler...)
      Dropthrough from fast to slow entry point is also implemented.
      Tailcalls are marked and converted into jumps at mangle time.
  12. 29 May, 2002 2 commits
  13. 28 May, 2002 2 commits
  14. 27 May, 2002 5 commits
    • simonpj's avatar
      [project @ 2002-05-27 16:13:42 by simonpj] · fd7d044f
      simonpj authored
      Make negative literals work in patterns
      The issue here is that
      	f (-1) = True
      	f x    = False
      should generate
      	f x = x == negate (fromInteger 1)
      rather than
      	f x = x == fromInteger (-1)
    • simonmar's avatar
      [project @ 2002-05-27 16:13:32 by simonmar] · 9aa4ed05
      simonmar authored
      Remove u2i, it was identical to getKey.
    • simonmar's avatar
      [project @ 2002-05-27 16:12:35 by simonmar] · e2505cab
      simonmar authored
      Use Unique.getKey instead of Unique.u2i (they are the same function).
    • simonpj's avatar
      [project @ 2002-05-27 15:28:07 by simonpj] · ef2b170c
      simonpj authored
      Allow infix type constructors
      This commit adds infix type constructors (but not yet class constructors).
      The documentation describes what should be the case.  Lots of tiresome
      changes, but nothing exciting.
      Allows infix type constructors everwhere a type can occur, and in a data
      or type synonym decl.  E.g.
      	data a :*: b = ....
      You can give fixity decls for type constructors, but the fixity decl
      applies both to the tycon and the corresponding data con.
    • simonpj's avatar
      [project @ 2002-05-27 10:20:20 by simonpj] · 4c5db78e
      simonpj authored
      Build TyDecl and ConDecl with the RdrHsSyn constructors, so the SysNames are right
  15. 24 May, 2002 1 commit
  16. 23 May, 2002 3 commits
  17. 21 May, 2002 1 commit
  18. 15 May, 2002 2 commits