[gentoo-lisp] [pxrist@gmail.com: Re: State of gnu common lisp]

[gentoo-lisp] [pxrist@gmail.com: Re: State of gnu common lisp]

[Thomas Kahle]

[-- Attachment #1: Type: text/plain, Size: 1169 bytes --]

Hi *,

I had a short email conversation with pchrist@g.o regarding
sci-mathematics/axiom, which I bumped to the latest version in the
science overlay.  The ebuild is suboptimal, but here is my original
message, any comments appreciated:

----snip----

over at the science overlay we are trying to revive
sci-mathematics/axiom which heavily relies on gnu common lisp (and no
other dialect).  In fact, the current ebuild will ship an internal
version of what axiom upstream calls gcl-2.6.8_pre4.

Do you have any specific plans of reviving gcl? What are the
difficulties that you are facing?

One problem that I found myself is that gcl (at least the version that
comes with axiom) seems to depend on <=dev-libs/gmp-4, so I emailed Camm
Maguire (who seems to be gcl upstream).  He was surprised to learn about
the release of gmp-5, which was not on his radar because appearently it
is not in Debian Unstable yet.

Ok, so this is about it.  I would love to see an unmasked gcl at some
point.  Let me know if I can help.  I'm currently being recruited as a
developer.

----- snap-----

Comments? Ideas?

Cheers,
Thomas

-- 
Thomas Kahle

[-- Attachment #2: Type: application/pgp-signature, Size: 316 bytes --]

Re: [gentoo-lisp] [pxrist@gmail.com: Re: State of gnu common lisp]

[Pascal J. Bourguignon]

On 2010/10/15, at 20:25 , Thomas Kahle wrote:

> Hi *,
>
> I had a short email conversation with pchrist@g.o regarding
> sci-mathematics/axiom, which I bumped to the latest version in the
> science overlay.  The ebuild is suboptimal, but here is my original
> message, any comments appreciated:
>
> ----snip----
>
> over at the science overlay we are trying to revive
> sci-mathematics/axiom which heavily relies on gnu common lisp (and no
> other dialect).  In fact, the current ebuild will ship an internal
> version of what axiom upstream calls gcl-2.6.8_pre4.
>
> Do you have any specific plans of reviving gcl? What are the
> difficulties that you are facing?
>
> One problem that I found myself is that gcl (at least the version that
> comes with axiom) seems to depend on <=dev-libs/gmp-4, so I emailed  
> Camm
> Maguire (who seems to be gcl upstream).  He was surprised to learn  
> about
> the release of gmp-5, which was not on his radar because appearently  
> it
> is not in Debian Unstable yet.
>
> Ok, so this is about it.  I would love to see an unmasked gcl at some
> point.  Let me know if I can help.  I'm currently being recruited as a
> developer.
>
> ----- snap-----
>
> Comments? Ideas?

Wouldn't it be better to port axiom to Common Lisp and let it run on  
any good CL implementation instead?  It would surely do nicely on SBCL.


-- 
__Pascal Bourguignon__
http://www.informatimago.com

Re: [gentoo-lisp] [pxrist@gmail.com: Re: State of gnu common lisp]

[Andrey G. Grozin]

On Fri, 15 Oct 2010, Thomas Kahle wrote:
> over at the science overlay we are trying to revive
> sci-mathematics/axiom which heavily relies on gnu common lisp (and no
> other dialect).  In fact, the current ebuild will ship an internal
> version of what axiom upstream calls gcl-2.6.8_pre4.
Why not just use fricas? It's in the main tree, and can be compiled by 
sbcl, cmucl, clozurecl, clisp, ecl (and probably gcl).

> I would love to see an unmasked gcl at some point.
Me too. Debian ships usable gcl for lots of architectures. I think we can 
just return to the old practice: take the 2.6.7 tarball + the (huge!) 
debian patch number <n>, and ship the result as 2.6.8_pre<n>.

Andrey

Re: [gentoo-lisp] [pxrist@gmail.com: Re: State of gnu common lisp]

[Thomas Kahle]

[-- Attachment #1: Type: text/plain, Size: 1330 bytes --]

On 10:04 Sat 16 Oct     , Andrey G. Grozin wrote:
> On Fri, 15 Oct 2010, Thomas Kahle wrote:
> > over at the science overlay we are trying to revive
> > sci-mathematics/axiom which heavily relies on gnu common lisp (and no
> > other dialect).  In fact, the current ebuild will ship an internal
> > version of what axiom upstream calls gcl-2.6.8_pre4.
> Why not just use fricas? It's in the main tree, and can be compiled by 
> sbcl, cmucl, clozurecl, clisp, ecl (and probably gcl).

Well, I did not know about it.  The other fork "OpenAxiom" is already in
the science overlay.  On the other hand, I also don't use axiom.  I'm
just trying to work on bugs assigned to sci-math@g.o.  In the end,
dropping support for the original axiom, given that there is fricas, is
one option to close these bugs.  

> > I would love to see an unmasked gcl at some point.
> Me too. Debian ships usable gcl for lots of architectures. I think we can 
> just return to the old practice: take the 2.6.7 tarball + the (huge!) 
> debian patch number <n>, and ship the result as 2.6.8_pre<n>.

Sounds good to me, and I think that is the setup axiom developers ship
because Camm Maguire is both: gcl upstream, and gcl's maintainer in
Debian.  He also provides the gcl versions for axiom.

Cheers,
Thomas



-- 
Thomas Kahle

[-- Attachment #2: Type: application/pgp-signature, Size: 316 bytes --]

Re: [gentoo-lisp] [pxrist@gmail.com: Re: State of gnu common lisp]

[Christian Faulhammer]

[-- Attachment #1: Type: text/plain, Size: 1753 bytes --]

Hi,

Thomas Kahle <tom111@gmx.de>:

> On 10:04 Sat 16 Oct     , Andrey G. Grozin wrote:
> > On Fri, 15 Oct 2010, Thomas Kahle wrote:
> > > over at the science overlay we are trying to revive
> > > sci-mathematics/axiom which heavily relies on gnu common lisp
> > > (and no other dialect).  In fact, the current ebuild will ship an
> > > internal version of what axiom upstream calls gcl-2.6.8_pre4.
> > Why not just use fricas? It's in the main tree, and can be compiled
> > by sbcl, cmucl, clozurecl, clisp, ecl (and probably gcl).
> 
> Well, I did not know about it.  The other fork "OpenAxiom" is already
> in the science overlay.  On the other hand, I also don't use axiom.
> I'm just trying to work on bugs assigned to sci-math@g.o.  In the end,
> dropping support for the original axiom, given that there is fricas,
> is one option to close these bugs.  

 And last rite axiom, with two alternatives available.

> > > I would love to see an unmasked gcl at some point.
> > Me too. Debian ships usable gcl for lots of architectures. I think
> > we can just return to the old practice: take the 2.6.7 tarball +
> > the (huge!) debian patch number <n>, and ship the result as
> > 2.6.8_pre<n>.
> 
> Sounds good to me, and I think that is the setup axiom developers ship
> because Camm Maguire is both: gcl upstream, and gcl's maintainer in
> Debian.  He also provides the gcl versions for axiom.

 That sounds like totally selfish and retarded.  Why not prepare a
proper release to ease burden for all the packagers of other
distributions out there.

V-Li

-- 
Christian Faulhammer, Gentoo Lisp project
<URL:http://www.gentoo.org/proj/en/lisp/>, #gentoo-lisp on FreeNode

<URL:http://gentoo.faulhammer.org/>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]