[Gcl-devel] Re: *features* - patch

[Michael Koehne]

Moin Camm Maguire,

> Greetings, and thanks so much for your work here!  I have one question
> -- do we even need CPU and OSTYPE keywords in the features list?  We
> don't have anything that hardware specific in GCL, do we?  Or is this
> simply conventionally expected?

  both:
  #+gcl simply conventionally expected - I dont expect a normal
    Debian/GNU/Linux system to claim wrong processor :MC68020
    or wrong OS :BSD386. The alternate would have been to drop
    every CPU and OSTYPE keyword from *features* in the other
    CPU/OSTYPE kombos in gcl/h/. I deceided not to drop them,
    but to fix them, to copy CLISP with its :PC386 feature to
    describe the architecture, and to resolve the CPU itself a
    bit deeper as :I386, :I486, :I586, :I586.

  #+clisp exports :FFI :PC386 :UNIX to tell about architecture.
    I therefore think about :BSD and :UNIX means ' it looks like
    :BSD and :UNIX ' and not that its realy UNIX (tm) AT&T or
    some real BSD derivate. I therefore left the :BSD and :UNIX
    references as they's been, outside the gcl/h/ config.guess.

  #+lucid the trampoline code that is using assembler optimisation
    is commented out - one might think about reenableing it, as
    Lucid was a free source company and is now history - but that
    not my scope.

  #+vaporware one might think about assembler optimisation in
    cline statements - would be realy easy, to do them now.

Bye Michael
-- 
  mailto:address@hidden             UNA:+.? 'CED+2+:::Linux:2.4.22'UNZ+1'
  http://www.xml-edifact.org/           CETERUM CENSEO WINDOWS ESSE DELENDAM