[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Gcl-devel] condition system proposal
From: |
Peter Wood |
Subject: |
Re: [Gcl-devel] condition system proposal |
Date: |
Tue, 5 Nov 2002 07:42:12 +0100 |
User-agent: |
Mutt/1.4i |
Hi,
On Mon, Nov 04, 2002 at 01:50:55PM +0300, Vadim V. Zhytnikov wrote:
> It seems that Peter is doing right things. Great!
>
> But I just wonder can we really bring conditions
> in sync with ANSI CL standard using the way
> GCL ansi image is build now with successive
> addition PCS as CLCS (conditions). If not
> then we sooner or later have to recreate
> build process from scratch.
>
This sounds right. I'd be really interested to hear what you think
this build process should look like (potentially).
> Second point is serror.lsp and SERROR
> package. As far as I understand this
> is old substitute for CL conditions.
> Can we get rid of this package as soon
> as ANSI conditions are in place?
> On the other hand I know that now
> Maxima relies on SERROR.
>
Yes. I think achieving ANSI compliance and maintaining backward
compatibility with packages that need a CLTL1 are going to prove to be
_incompatible_ goals. I'm afraid we need to decide what we want to
do: maintain a CLTL1, or make an ANSI Common LIsp (or both -
seperately).
Camm, you have said that maintaining compatibilty with Maxima/ACL2 is
a requirement (along with support for the 11 Debian architecures).
Would you be amenable to an 'experimental' branch which aims at strict
compliance and is allowed to have a narrower focus? I know you have
said before that you would prefer to avoid this.
Regards,
Peter