axiom-developer
[Top][All Lists]
Advanced

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

RE: [Axiom-developer] Old Lisp


From: Bill Page
Subject: RE: [Axiom-developer] Old Lisp
Date: Tue, 15 Aug 2006 00:45:36 -0400

Camm,

When you have a moment could you please confirm that I am not
leading Gaby astray here. :-)

Thanks.

On August 12, 2006 10:34 PM Gaby wrote:
> 
> "Bill Page" <address@hidden> writes:
> 
> | On August 12, 2006 8:56 PM Gaby wrote:
> | > ... 
> | > There are still some patches against GCL that Tim developed.
> | > Ideally, I think we should not be in the business of patching
> | > GCL. Camm has been collaborating with us very effectively; we
> | > should communicate those issues with him and see how they can
> | > be addressed in GCL upstream.
> | > 
> | 
> | I think all of the patches made by Tim are obsolete. I agree
> | that we should not be in the business of patching GCL. Although
> | GCL is currently the primary vehicle for delivering Axiom, it
> | is not the only Lisp that is capable of doing so. (For example
> | there is the open source CMUCL version of Axiom done by Jurgen
> | Weiss and the commercial version of Axiom used yet another lisp
> | and that lisp is now open source too.
> 
> OK.
> 
> | In fact I am fairly sure that Camm has already addressed all of
> | the GCL issues that have been identified so far. None of these
> | patches are used in the Debian release of Axiom which is built
> | from an unmodified standard distribution of GCL. There is no
> | reason that I can think of way this should not become the standard
> | way of building Axiom from GCL on all platforms (including Windows).
> 
> For GCL-2.6.7, Tim has a patch against configure.in.  The reason
> being that:
> 
>    There is a typo in configure.in that is only detected under
>    some versions of bash. The problem is a missing single-quote
>    mark.
> 
> For that to work correctly, there is an implicit assumption that
> the user's build environment has Autoconf -- to regenerate configure
> from the patched configure.in.  I don't how serious that is.

It was corrected in GCL-2.6.7 shortly after Tim took the tarball
image from cvs.

> Anyway, it GCL-2.6.7 that we might just abandon from the silver
> branch.

Good idea.

> 
> For GCL-2.6.8pre we have four patches:
> 
>   * gcl-2.6.8pre.socket.patch
>   * gcl-2.6.8pre.libspad.patch

The above two are handled by compiler-link in an unmodified
previously installed copy of GCL.

>   * gcl-2.6.8pre.toploop.patch

This one is handled by setting an option.

>   * gcl-2.6.8pre.collectfn.fix

This one just copies (and renames) gcl_collectfn.lsp to a different
location in the file system and copies sys-proclaim.lisp to that
location for use at a later stage in the build. It has nothing to
do with patching GCL and should be addressed by modifications to
the appropriate makefile.pamphlet.

> 
> Can you check with Camm that they are indeed not needed anymore?
> I'm sorry, I don't fully understand the issues they are trying to
> address.
> 

Regards,
Bill Page.






reply via email to

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