1. 25 Feb, 2000 4 commits
  2. 24 Feb, 2000 8 commits
  3. 23 Feb, 2000 1 commit
    • lewie's avatar
      [project @ 2000-02-23 19:41:50 by lewie] · e87d56ce
      lewie authored
      Handle `with' more cleverly.  I was generating partially applied methods
      for the case where the `with' expression was also overloaded, but this
      was buggy, and completely unnecessary.  Instead, simply force the method
      binding at the point of the `with' expression (we reap no benefits from
      pushing the sharing further out anyway), and release the remainder of
      the method's context into the LIE.
  4. 22 Feb, 2000 3 commits
    • panne's avatar
      [project @ 2000-02-22 15:47:56 by panne] · b78eb7be
      panne authored
      Load deprecations from interface files into a deprecation environment
      which maps Names to RenamedDeprecations. This map is not used yet, but
      very soon it will...
      This commit fixes a bug related to implicit parameters, too:
      Previously, an interface file containing the name "with" could not be
      read by the interface parser. This broke Malcolm's HaXml 0.9 (released
      today). Remember Sven's glaexts-commandment (Jeffrey? :-) : Always
      keep Lex.lhs's ghcExtensionKeywordsFM and ParseIface.y's var_fs
      production in synch!
    • panne's avatar
      [project @ 2000-02-22 12:45:30 by panne] · c01dc71d
      panne authored
      Space leak paranoia: Don't keep whole InterfaceDetails alive
      throughout the middle part of the compiler
    • simonmar's avatar
      [project @ 2000-02-22 12:09:23 by simonmar] · b310c854
      simonmar authored
      Ignore SIGPIPE, since our I/O library handles EPIPE properly and
      SIGPIPE tends to cause the Haskell program to exit silently and
      Found-by: furiously pressing the refresh button in a browser connected
      to my Haskell web server :-)
  5. 21 Feb, 2000 3 commits
    • panne's avatar
      [project @ 2000-02-21 19:06:11 by panne] · 8486d65e
      panne authored
      Extended interface mangler with deprecations handling.
      NOTE: Whoever wrote this part of the driver without adding a HUGE
      warning in MkIface.lhs to keep both files in synch deserves the most
      severe punishment imaginable (huge stick, etc.)!  >:-(  The seemingly
      magical random rearrangement of the interface file outside Haskell's
      realm can drive you mad if you forget this...
    • panne's avatar
      [project @ 2000-02-21 18:56:02 by panne] · 105fb636
      panne authored
      Fixed typo in deprecations parsing
    • panne's avatar
      [project @ 2000-02-21 18:55:19 by panne] · 3909a13c
      panne authored
      Write deprecations into interface files.
  6. 20 Feb, 2000 1 commit
  7. 18 Feb, 2000 4 commits
  8. 17 Feb, 2000 4 commits
  9. 16 Feb, 2000 2 commits
  10. 15 Feb, 2000 4 commits
    • panne's avatar
      [project @ 2000-02-15 22:18:16 by panne] · 06619533
      panne authored
      First steps towards DEPRECATED before Rosebank (12yrs) takes its toll.
      Nothing very functional yet, but at least hsc can be compiled and it
      still compiles the Prelude.
      Parsing the pragma turned out to be a little bit more complicated than
      expected, here the comment from Parser.y:
         The place for module deprecation is really too restrictive, but if it
         was allowed at its natural place just before 'module', we get an ugly
         s/r conflict with the second alternative. Another solution would be the
         introduction of a new pragma DEPRECATED_MODULE, but this is not very nice,
         either, and DEPRECATED is only expected to be used by people who really
         know what they are doing. :-)
      Net result: Module deprecation is allowed exactly behind the module's
      name and nowhere else. I probably have to think a little bit more
      about this some day...
    • sewardj's avatar
      [project @ 2000-02-15 15:14:09 by sewardj] · 2c64208b
      sewardj authored
      Implement CONST_INT_big and CONST_ADDR_big.
    • sewardj's avatar
      [project @ 2000-02-15 13:16:19 by sewardj] · 7c1668b4
      sewardj authored
      Backend interop fixes:
      -- Make Hugs use the same constructor tag numbering as GHC, viz, starting
         at zero.
      -- Evaluator.c: when unwinding the stack on entering a constructor,
         return to the scheduler if a RET_{VEC_}{SMALL|BIG} is found on the
    • sewardj's avatar
      [project @ 2000-02-15 11:24:20 by sewardj] · f460a89b
      sewardj authored
      In hugsprimRunIO_toplevel, use catchException instead of catch.
      At least exceptions can then be caught, although Hugs then goes on
      to die with an assertion failure in unblockAsyncExceptionszh_ret_entry,
      probably caused by wrongly entering it twice in immediate succession.
  11. 14 Feb, 2000 6 commits