MAXIMA, LISP and COMMON-LISP package on GCL

James Amundson ÐÉÛÅÔ:

> On Thu, 2004-04-22 at 02:29, Vadim V. Zhytnikov wrote:
> 
> 
>>On the other hand ANSI GCL has two LISP and
>>COMMON-LISP packages.  The former is legacy CLtL1
>>package and the latter is intended to be ANSI
>>compliant.  Thus ANSI features remains unexposed
>>in MAXIMA package.  This is the reason why new
>>James' to_lisp() REPL doesn't handle lisp errors
>>properly on GCL (maxima::*debugger-hook* isn't
>>the same symbol as cl::*debugger-hook*).
> 
> 
> Vadim, you must be reading my mind lately. You have fixed and/or
> identified the sources of all the problems I have been most worried
> about. I am very grateful.
> 
> 
>>Ultimately on can get rid of old LISP package by
>>renaming it to LISP-LEGACY and making LISP
>>alias to COMMON-LISP.  Unfortunately it can't be done
>>since GCL Maxima makes use of another old style SERROR
>>package which conflicts with COMMON-LISP (conflicts with
>>CONDITIONS which supersedes SERROR in COMMON-LISP).
>>
>>The quick fix to the problem is to import
>>required ANSI symbols from COMMON-LISP explicitly.
> 
> 
> Yes, lets do that.
> 
> 
>>This is not so hard since right now we need just few of
>>them.  But it isn't right approach in the long run.
>>IMHO the right way is to remove SERROR entirely
>>but this is not so simple and require thorough
>>testing. I think it can be done later with next
>>Maxima and GCL releases.
> 
> 
> I agree. 
> 
> --Jim
> 

What do your have in your "must be done or fixed before release" list?
Maybe I or someone else could help to speed thing up?
I just want to remind about automake 1.8 incompatibility and
default value for cmucl runtime.  Unfortunately both items are
beyond my automake/autoconf competence.


-- 
      Vadim V. Zhytnikov

       <vvzhy@mail.ru>
      <vvzhy@netorn.ru>