[Gcl-devel] Re: your mail

[Camm Maguire]

Greetings!

Tuukka Toivonen <address@hidden> writes:

> On Wed, 19 Dec 2001, Camm Maguire wrote:
> 
> >Greetings, and thank you so much for submitting this patch!  I'm
> 
> Thank you for maintaining the GCL--and my fresh Debian operating
> system installation. :)
> 
> >is optionally performed at runtime with (si::init-readline),
> 
> I assume that you have not modified my patch significantly, ie.
> the LISP calling convention is the same.
> 
> I'm a bit worried about the API stability: instead of a four
> argument initialization routine like this:
>    (si::READLINE-INIT t "Gcl" 1 '("first" "second" "third" "rest"))
> I would rather have less mandatory arguments and instead use
> optional arguments, that have default values if not given. This
> is possible with C++, and I know it is also possible with LISP,
> but I don't know how to read optional arguments from Lisp in the C.
> 
> However, if you apply the patch now, it will be more difficult
> to change the API later. But I don't think I can myself alone
> change the API better, unless helped by someone more knowledgeable.
> 
> I sent a question to Dr. Schelter, but unfortunately, it was too late.
> 

OK, you might want to take a look at what I did with the patch.  It is
initialized with a (si::init-readline) which basically defaults the
arguments to readline-init.  Feedback most appreciated!

> >1) I'm a bit new to this, but it appears that lisp in general and gcl
> >in particular cold have difficulties linking with third party code
> >potentially calling malloc.  I know Dr. Schelter had to replace with
> 
> I never had problems with this. I cannot comment, except that
> "works for me" (with glibc 2.0.6 primarily).
> 
> >alloca in gmp, for example.  Is this only because the bignums were
> 
> If stack works, why not alloca? I'm completely clueless here.
> 
> >   keyword list for maxima.  I believe one could just slightly modify
> >   this simple lisp function you supplied in init_gcl.lsp, which now
> >   is invoked at runtime, and then have maxima keyword completion.
> 
> Yes, that is the intention. Just insert the similar code in some
> Maxima initialization routine. In fact, I think I have already made a
> patch for that. I'll dig it up for you, tomorrow.

I'd really like to understand how to best get this working for
maxima.  Suggestions from the list?


> 
> Please tell me if you think I can help you any way, althought
> I'm so clueless with Lisp. :(
> 
> 
> 

Take care,

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