[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gcl-devel] New lisp code

From: Camm Maguire
Subject: [Gcl-devel] New lisp code
Date: 24 Jan 2002 14:43:15 -0500

Greetings!  Just committed R. Toy's port of 'defsystem' to gcl 2.5.0.
This brings up a few items:

1) What procedures do we want to put in place regarding the adoption
   of 3rd party lisp code?  We obviously have to worry about the
   license, for example.  Do we also need some independent
   verification of the functionality?  What about performance?  Should
   we write our own in some cases?  Should we leave in the #+ and #-?
   Even if it is permissible, is it polite to incorporate code from
   other lisp systems?  

   As to the last question, my opinion is that all gcl has to offer is
   its performance/portability stemming from its compiler to C code,
   and perhaps key routine implementation directly into C.  We don't
   now, and shouldn't at least in the near future, hope to offer nifty
   items on the lisp end -- in fact, we need to play catch-up here.
   So I do not feel we are compromising any of gcl's unique strengths
   by incorporating external lisp code where permissible.  Other
   opinions welcome!

2) For those interested, the tech procedure is quite simple:
        a) install lisp source as ???.lsp in lsp dir.
        b) fire up saved_gcl
        c) (compile-file "../lsp/???.lsp" :c-file t :h-file t
                :data-file t)
        d) edit produced .c file, replace inlined header defines with 
                '#include <cmpinclude.h>', rename init_code to
        e) add .o targets to lsp/makefile and unixport/makefile
        f) edit unixport/saved_gcl.c, and add line like

3) Still have problems running gcl with ilisp:
          ;;; Loading /usr/share/emacs20/site-lisp/ilisp/ilisp-pkg.lisp
          Error: DEFPACKAGE::OPTION is invalid as a function.

4) What are we going to do about the ansi-loop patch also submitted by
   R. Toy?  I notice that maxima uses both (loop and (sloop.  Won't we
   have to protect the new loop with a runtime flag?

Take care, 
Camm Maguire                                            address@hidden
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah

reply via email to

[Prev in Thread] Current Thread [Next in Thread]