Hello all,
I just get GCL-2.4.1 installed and try to use it. Unluckily I have
trouble with it. The statement below
(DEFCLASS New-Class () ()) ;; will be redefined later.
caused error messages as follows:
============================================================
Error: The function DEFCLASS is undefined.
Fast links are on: do (si::use-fast-links nil) for debugging
Error signalled by EVAL.
Broken at EVAL. Type :H for Help.
============================================================
I guess I might miss something important before the statement. How can
I use the `DEFCLASS' correctly? I would apologize for this naive
question.
Thanks in advance.
--
Zhongtao Zhu Tel: +86 10 6278 2266
Department of Computer, Tsinghua University, Beijing, China
Zhongtao Zhu <········@tsinghua.edu.cn> wrote in message news:<··············@localhost.localdomain>...
> Hello all,
>
> I just get GCL-2.4.1 installed and try to use it. Unluckily I have
> trouble with it. The statement below
>
> (DEFCLASS New-Class () ()) ;; will be redefined later.
GCL is not a fully conforming implementation of Common Lisp. Among the
many features missing is the object system, which can be installed as
a separate add-on based on PCL. (This might have changed; I don't know
whether 2.4.1 is the most recent version; I've been reading about some
new activity on the GCL development front).
> I guess I might miss something important before the statement. How can
> I use the `DEFCLASS' correctly? I would apologize for this naive
> question.
There are other free implementations of Lisp; it might be a good idea
to look at those.
···@ashi.footprints.net (Kaz Kylheku) writes:
> Zhongtao Zhu <········@tsinghua.edu.cn> wrote
> > Hello all,
> >
> > I just get GCL-2.4.1 installed and try to use it. Unluckily I have
> > trouble with it. The statement below
> >
> > (DEFCLASS New-Class () ()) ;; will be redefined later.
>
> GCL is not a fully conforming implementation of Common Lisp. Among the
> many features missing is the object system, which can be installed as
> a separate add-on based on PCL. (This might have changed; I don't know
> whether 2.4.1 is the most recent version; I've been reading about some
> new activity on the GCL development front).
>
> > I guess I might miss something important before the statement. How can
> > I use the `DEFCLASS' correctly? I would apologize for this naive
> > question.
>
> There are other free implementations of Lisp; it might be a good idea
> to look at those.
Thank you all, dear friends here in the list,
I switched to AllegroCL 6.2 and it works through. Now everything is on
its track. By the way, I use a RedHat 8.0 distro which was just
released several days ago. I believe my experience about using the
latest AllegroCL smoothly on the newest RedHat distro may be helpful
for the people who are always chase with the latest versions.
Best regards,
--
Zhongtao Zhu Tel: +86 10 6278 2266
Department of Computer, Tsinghua University, Beijing, China
···@ashi.footprints.net (Kaz Kylheku) writes:
> Zhongtao Zhu <········@tsinghua.edu.cn> wrote
> > Hello all,
> >
> > I just get GCL-2.4.1 installed and try to use it. Unluckily I have
> > trouble with it. The statement below
> >
> > (DEFCLASS New-Class () ()) ;; will be redefined later.
>
> GCL is not a fully conforming implementation of Common Lisp. Among the
> many features missing is the object system, which can be installed as
> a separate add-on based on PCL. (This might have changed; I don't know
> whether 2.4.1 is the most recent version; I've been reading about some
> new activity on the GCL development front).
>
> > I guess I might miss something important before the statement. How can
> > I use the `DEFCLASS' correctly? I would apologize for this naive
> > question.
>
> There are other free implementations of Lisp; it might be a good idea
> to look at those.
Thank you all, dear friends here in the list,
I switched to AllegroCL 6.2 and it works through. Now everything is on
its track. By the way, I use a RedHat 8.0 distro which was just
released several days ago. I believe my experience about using the
latest AllegroCL smoothly on the newest RedHat distro may be helpful
for the people who always chase with the latest versions.
Best regards,
--
Zhongtao Zhu Tel: +86 10 6278 2266
Department of Computer, Tsinghua University, Beijing, China
Zhongtao Zhu wrote:
> I switched to AllegroCL 6.2 and it works through. Now everything is on
> its track. By the way, I use a RedHat 8.0 distro which was just
> released several days ago. I believe my experience about using the
> latest AllegroCL smoothly on the newest RedHat distro may be helpful
> for the people who always chase with the latest versions.
Well, so share :-). What's up?
I finally switched from KDE to Gnome because the OpenMotif libraries used by
KDE conflicted with the lesstif libraries needed by LispWorks Personal
Edition and I didn't want to futz with trying to extract individual .so
files from the lesstif RPM's to work around it. BTW, does anyone have any
tips (besides coding directly to X, that is) on how to write a portable
Linux application having a GUI? I would imaging that people running KDE on
RH 8.0 are going to have problems with apps built with LispWorks unless
their CLIM primitives are coded at the X level. Even better, could someone
from Xanalys comment on lesstif/OpenMotif conflicts under RH 8 and what
they're telling customers to do (beside staying on 7.3 :-)?
faa
Thanks for everyone's feedback. As to the questions...
I don't know if Open Motif is an absolute need for KDE 3.1. It does seem to
be listed as a dependency for the kdebase library RPM in RedHat 8.0.
I tried to run LW Personal Edition from the tarball on a stock RH 8.0, with
whatever version of Open Motif they use installed and without lesstif. I
got errors somewhere in the C API wrapper layers. I don't remember the
exact version numbers of everything, but I can dig them up. I even went so
far as to try to compile a new version of lesstif before I gave up. In any
case, it's quite possible that RedHat might not have a quite compatible
version of either the lesstif library, the OM libraries, or the development
libraries for any of these. In any case, getting rid of KDE and Open Motif
and installing lesstif fixed my problems.
And again, thanks for the info and for the version numbers. I'll report
back as I have time and results.
faa
"Frank A. Adrian" <·······@ancar.org> writes:
> I tried to run LW Personal Edition from the tarball on a stock RH
> 8.0, with whatever version of Open Motif they use installed and
> without lesstif. I got errors somewhere in the C API wrapper
> layers. I don't remember the exact version numbers of everything,
> but I can dig them up. I even went so far as to try to compile a
> new version of lesstif before I gave up. In any case, it's quite
> possible that RedHat might not have a quite compatible version of
> either the lesstif library, the OM libraries, or the development
> libraries for any of these. In any case, getting rid of KDE and
> Open Motif and installing lesstif fixed my problems.
For what it's worth here's a bug report I sent to Xanalys and SuSE
about a year ago.
Edi> Today I updated my laptop's Linux installation to SuSE 7.3. Now
Edi> LispWorks (Personal Edition, 4.1.20) doesn't work anymore. If I
Edi> start it, I only get the following message:
Edi> LispWorks: The Common Lisp Programming Environment Personal Edition
Edi> Copyright (C) 1987-2001 Xanalys Incorporated. All rights reserved.
Edi> Version 4.1.20
Edi> Saved by Xanalys as lispworks-personal-4120, at 23 Mar 2001 14:40
Edi> User edi on bird
Edi> Warning: Unknown KeySym value: 0x20AC
Edi> Warning: Unknown KeySym value: 0xFE60
Edi> Warning: (Xt Warning) Representation size 2 must match superclass's to override
Edi> marginWidth
Edi> Warning: (Xt Warning) Representation size 1 must match superclass's to override
Edi> resizePolicyError during GUI startup:
Edi> Signal : segmentation violation(11)
Edi> This is with SuSE's OpenMotif, their version is 2.1.30MLI4-134,
Edi> Sun Sep 23 22:35:18 2001. If I switch to LessTif, the problem
Edi> is gone - the program starts fine, albeit still with the
Edi> message
Edi> Warning: Unknown KeySym value: 0x20AC
Edi> Warning: Unknown KeySym value: 0xFE60
Edi> (which also was there with SuSE 7.2).
Edi> However, LessTif has a nasty bug concerning the placement of
Edi> contextual menus (which I reported earlier), so I'd prefer to
Edi> use OpenMotif instead. The combination of LispWorks and
Edi> OpenMotif worked fine with SuSE 7.2.
With the help of Xanalys support I was able to solve the problem by
"downgrading" from the SuSE 7.3 RPM of OpenMotif (2.1.30MLI4-134) to
their 7.2 version (2.1.30MLI4-85, same version but different
builds). SuSE support refused to admit that there might be a problem
with their libraries.
Maybe your problem is related to this one?
Edi.
"Frank A. Adrian" <·······@ancar.org> writes:
> I finally switched from KDE to Gnome because the OpenMotif libraries used by
> KDE conflicted with the lesstif libraries needed by LispWorks Personal
> Edition and I didn't want to futz with trying to extract individual .so
Are you sure LW needs lesstif? 4.1.20 works here with Open Motif.
Andras
······@math.bme.hu (Simon Andr�s) writes:
> "Frank A. Adrian" <·······@ancar.org> writes:
>
>> I finally switched from KDE to Gnome because the OpenMotif libraries used by
>> KDE conflicted with the lesstif libraries needed by LispWorks Personal
>> Edition and I didn't want to futz with trying to extract individual .so
>
> Are you sure LW needs lesstif? 4.1.20 works here with Open Motif.
And are you sure KDE needs OpenMotif?
"Frank A. Adrian" <·······@ancar.org> writes:
> I finally switched from KDE to Gnome because the OpenMotif libraries
> used by KDE conflicted with the lesstif libraries needed by
> LispWorks Personal Edition and I didn't want to futz with trying to
> extract individual .so files from the lesstif RPM's to work around
> it. BTW, does anyone have any tips (besides coding directly to X,
> that is) on how to write a portable Linux application having a GUI?
> I would imaging that people running KDE on RH 8.0 are going to have
> problems with apps built with LispWorks unless their CLIM primitives
> are coded at the X level. Even better, could someone from Xanalys
> comment on lesstif/OpenMotif conflicts under RH 8 and what they're
> telling customers to do (beside staying on 7.3 :-)?
I'm using LW professional 4.2.x on Linux (Gentoo 1.1a) with OpenMotif
and KDE without any problems. I also had no problems with this
combination when using LW personal 4.1 on SuSE 7.3/8.0.
In fact, now I remember that I had some annoying little bugs in the
IDE (like the context menu being far away from the mouse cursor) when
using Lesstif which magically disappeared after I switched to
OpenMotif.
Edi.
From: Erik Naggum
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <3242812738780708@naggum.no>
* Zhongtao Zhu
| I just get GCL-2.4.1 installed and try to use it. Unluckily I have
| trouble with it.
GCL is still a very old Common Lisp and does not purport to conform (as
the formal language goes) to the ANSI standard for Common Lisp. It is a
Common Lisp that refers to the 1984 book Common Lisp the Language. Books
on Lisp from the same time period will be most useful with GCL.
| I guess I might miss something important before the statement. How can I
| use the `DEFCLASS' correctly? I would apologize for this naive question.
Use any of the other free Common Lisp implementations. Since you have
just installed it, it should not be a loss to deinstall it and install
the Trial Edition of Allegro CL from Franz Inc <www.franz.com>, or the
Personal Edition of LispWorks from Xanalys <www.xanalys.com>, or any of
the supposedly free, but generally having the same kind of restrictions
as the above free offerings from commercial vendors, implementations
CMUCL, SBCL, or CLISP. Depending on your platform, there are also
Macintosh Common Lisp, Corwin Common Lisp, and several others.
See <www.lisp.org> for more information.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Paul F. Dietz
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <3D9EF56F.3070203@dls.net>
Erik Naggum wrote:
> * Zhongtao Zhu
> | I just get GCL-2.4.1 installed and try to use it. Unluckily I have
> | trouble with it.
>
> GCL is still a very old Common Lisp and does not purport to conform (as
> the formal language goes) to the ANSI standard for Common Lisp. It is a
> Common Lisp that refers to the 1984 book Common Lisp the Language. Books
> on Lisp from the same time period will be most useful with GCL.
You can download the source to gcl and configure it with the --enable-ansi
flag, then recompile. It will then be closer to ANSI CL. DEFCLASS, at least,
will be defined.
http://savannah.gnu.org/cvs/?group=gcl
Paul
"Paul F. Dietz" <·····@dls.net> writes:
> Erik Naggum wrote:
> > * Zhongtao Zhu
> > | I just get GCL-2.4.1 installed and try to use it. Unluckily I have
> > | trouble with it.
> > GCL is still a very old Common Lisp and does not purport to
> > conform (as
> > the formal language goes) to the ANSI standard for Common Lisp. It is a
> > Common Lisp that refers to the 1984 book Common Lisp the Language. Books
> > on Lisp from the same time period will be most useful with GCL.
>
> You can download the source to gcl and configure it with the --enable-ansi
> flag, then recompile. It will then be closer to ANSI CL. DEFCLASS, at least,
> will be defined.
>
> http://savannah.gnu.org/cvs/?group=gcl
Why isn't --enable-ansi true by default?
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
715 Broadway 10th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
From: Paul F. Dietz
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <3D9F5B45.1050502@dls.net>
Marco Antoniotti wrote:
> Why isn't --enable-ansi true by default?
I'm not entirely sure. I think it may be because gcl was mostly
being used for maxima and acl2, and because it's not yet really
ANSI CL even with --enable-ansi.
Paul
Greetings!
Just a brief note on the state of GCL -- ansi support is still
preliminary, and therefore not enabled by default. That having been
said, we've been making some significant progress toward ansi
compliance, and hope to be in reasonable shape in 3-6 months. To give
you a rough idea, we envision an ansi-test suite of eventually some
10,000 tests. Currently we have a suite of ~4200, of which ~4000
pass. Any interested volunteers are welcome to assist, of course.
While the latest stable release is 2.4.3, basically a stabilization of
Dr. Schelter's work, CVS is generally kept in good shape. CVS
snapshot binary packages for various architectures are kept in the cvs
subdirectory of the GCL ftp site.
GCL has currently been ported to Windows and all 11 of the Debian
architectures, where it successfully builds the Debian maxima package
including all maxima regression tests. ACL2 also ships a GCL image,
and the forthcoming axiom is supposed to have a GCL target as well.
Take care,
Marco Antoniotti <·······@cs.nyu.edu> writes:
> "Paul F. Dietz" <·····@dls.net> writes:
>
> > You can download the source to gcl and configure it with the --enable-ansi
> > flag, then recompile. It will then be closer to ANSI CL. DEFCLASS, at
> > least, will be defined.
> >
> > http://savannah.gnu.org/cvs/?group=gcl
>
> Why isn't --enable-ansi true by default?
>
> Cheers
>
> --
> Marco Antoniotti ========================================================
> NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
> 715 Broadway 10th Floor fax +1 - 212 - 995 4122
> New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
> "Hello New York! We'll do what we can!"
> Bill Murray in `Ghostbusters'.
--
Camm Maguire ····@enhanced.com
==========================================================================
"The earth is but one country, and mankind its citizens." -- Baha'u'llah
>Camm Maguire <····@enhanced.com> wrote in message
···················@intech19.enhanced.com
[...]
> GCL has currently been ported to Windows and all 11 of the Debian
> architectures, where it successfully builds the Debian maxima package
> including all maxima regression tests. ACL2 also ships a GCL image,
> and the forthcoming axiom is supposed to have a GCL target as well.
Hi, is there any built GCL binary that can be download? Because, I got many
problems when I tried to build GCL 2.4.3 on MinGW on Win98... Or, someone
knows where to find helpful information that could help me to build GCL
correctly on MinGW?
Thank you!
Joe
From: Erik Naggum
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <3242816926170269@naggum.no>
* Erik Naggum
| Depending on your platform, there are also Macintosh Common Lisp,
| /Corwin Common Lisp/, and several others.
That would be Corman Lisp, <www.cormanlisp.com>.
(Thanks for the mailed correction.)
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> writes:
[...]
> Use any of the other free Common Lisp implementations. Since you have
> just installed it, it should not be a loss to deinstall it and install
> the Trial Edition of Allegro CL from Franz Inc <www.franz.com>, or the
> Personal Edition of LispWorks from Xanalys <www.xanalys.com>, or any of
> the supposedly free, but generally having the same kind of restrictions
> as the above free offerings from commercial vendors, implementations
> CMUCL, SBCL, or CLISP. Depending on your platform, there are also
> Macintosh Common Lisp, Corwin Common Lisp, and several others.
"Supposedly free" seems like a loaded phrase. CLISP does appear to be
a bit problematic for someone wanting to sell proprietary LISP
programs, but as far as I can tell, CMUCL and SBCL are available for
just about any use imaginable, being restricted only by the BSD
advertising clause on some components (the rest being public domain).
The free offerings from Franz Inc and Xanalys have licenses clearly
intended to prevent their use for production of proprietary code
(quite reasonably). The LispWorks free trial even has a limited heap,
which might limit its use in producing open source/personal software.
> See <www.lisp.org> for more information.
Yes, great resource.
From: Erik Naggum
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <3242823151506921@naggum.no>
* Bruce Stephens
| "Supposedly free" seems like a loaded phrase.
Somebody was sure to pick up on that. Yes, but it is because "free" is a
loaded term to begin with. The freer it appears, the more hidden its
costs. TANSTAAFL. However, it can sometimes take quite some time before
you realize what you have had to pay and some people never pay it. This
is why it is important to think about the costs up front.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Bruce Stephens <············@cenderis.demon.co.uk> wrote:
> The LispWorks free trial even has a limited heap, which might limit
> its use in producing open source/personal software.
Comparing the Windows versions (which is what I'm currently looking
at), the LispWorks heap limit is about three times as large as what
the free offering from Franz is allowing.
What I'd probably find more uncomfortable about LispWorks Personal
Edition if I wanted to use the free version for serious development is
the session time limit. You have to re-start the environment every
few hours.
--
Marcus Breiing
···················@breiing.com (will expire)
From: Lieven Marchand
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <874rc1otx9.fsf@wyrd.be>
Erik Naggum <····@naggum.no> writes:
> Depending on your platform, there are also Macintosh Common Lisp,
> Corwin Common Lisp, and several others.
^^^^^^ Corman
Have you been reading Zelazny lately, Erik?
--
Hai koe, zei de stier,
Kom mee met mij in de wei,
Dan zijn we tweezaam.
Lieven Marchand <···@wyrd.be>
Centuries ago, Nostradamus foresaw when Zhongtao Zhu <········@tsinghua.edu.cn> would write:
> Hello all,
>
> I just get GCL-2.4.1 installed and try to use it. Unluckily I have
> trouble with it. The statement below
>
> (DEFCLASS New-Class () ()) ;; will be redefined later.
>
> caused error messages as follows:
> ============================================================
> Error: The function DEFCLASS is undefined.
> Fast links are on: do (si::use-fast-links nil) for debugging
> Error signalled by EVAL.
> Broken at EVAL. Type :H for Help.
> ============================================================
>
> I guess I might miss something important before the statement. How can
> I use the `DEFCLASS' correctly? I would apologize for this naive
> question.
In this case, "using DEFCLASS correctly" involves finding another Lisp
implementation.
GCL is, more or less, an implementation of what Common Lisp was in the
days of the /first/ edition of "Common Lisp, The Language."
A number of areas of Common Lisp had not settled down to the point of,
well, "existing." CLOS is one example; the LOOP facility is another.
You would probably be /much/ better served by looking for something
that tries to implement ANSI Common Lisp. There are quite a number of
"free" options, including CLISP, CMU-CL, SBCL, CLiCC, ECoLisp, and
ECLS, all of which are more likely to include more of CLOS.
There apparently have been some recent efforts going into bringing GCL
more up-to-date; I somehow suspect they might be better spent on
ECoLisp or ECLS (which have a similar "compile the CL into C first"
approach as found in GCL).
You might want to indicate some info on what platform you are using,
and what you're generally trying to do, as that might help guide
suggestions...
--
(reverse (concatenate 'string ··········@" "enworbbc"))
http://www.ntlug.org/~cbbrowne/commonlisp.html
Rules of the Evil Overlord #145. "My dungeon cell decor will not
feature exposed pipes. While they add to the gloomy atmosphere, they
are good conductors of vibrations and a lot of prisoners know Morse
code." <http://www.eviloverlord.com/>
From: Christopher C. Stacy
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <u65wgld6n.fsf@dtpq.com>
>>>>> On 5 Oct 2002 19:46:11 GMT, Christopher Browne ("Christopher") writes:
Christopher> Centuries ago, Nostradamus foresaw when Zhongtao Zhu <········@tsinghua.edu.cn> would write:
>> Hello all,
>>
>> I just get GCL-2.4.1 installed and try to use it. Unluckily I have
>> trouble with it. The statement below
>>
>> (DEFCLASS New-Class () ()) ;; will be redefined later.
>>
>> caused error messages as follows:
>> ============================================================
>> Error: The function DEFCLASS is undefined.
>> Fast links are on: do (si::use-fast-links nil) for debugging
>> Error signalled by EVAL.
>> Broken at EVAL. Type :H for Help.
>> ============================================================
>>
>> I guess I might miss something important before the statement. How can
>> I use the `DEFCLASS' correctly? I would apologize for this naive
>> question.
Christopher> In this case, "using DEFCLASS correctly" involves finding another Lisp
Christopher> implementation.
Christopher> GCL is, more or less, an implementation of what Common Lisp was in the
Christopher> days of the /first/ edition of "Common Lisp, The Language."
Christopher> A number of areas of Common Lisp had not settled down to the point of,
Christopher> well, "existing." CLOS is one example; the LOOP facility is another.
LOOP was brought into Common Lisp from MACLISP, and I am sure
there must be a LOOP library implementation that can be optionally
loaded into GCL. LOOP hadn't made it into the Common Lisp standard,
but it certainly existed in all its full-blown glory.
Also back in those days, CLOS was faked in implementations that didn't
have it yet by using a library called PCL (Portable Common LOOPS).
The name is confusing: it had nothing to do with LOOP.
I think it stood for "Lisp Object Oriented System",
and was based on what came out of Xerox.
From: Paul F. Dietz
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <3D9F777D.3040100@dls.net>
Christopher C. Stacy wrote:
> LOOP was brought into Common Lisp from MACLISP, and I am sure
> there must be a LOOP library implementation that can be optionally
> loaded into GCL. LOOP hadn't made it into the Common Lisp standard,
> but it certainly existed in all its full-blown glory.
Configure gcl with --enable-ansi and you get LOOP. (I would not
be surprised if most lisps' LOOP implements were descended from
MIT's.)
> Also back in those days, CLOS was faked in implementations that didn't
> have it yet by using a library called PCL (Portable Common LOOPS).
And this, too. CMU CL also uses a (now increasingly modified)
version of PCL.
Paul
From: Erik Naggum
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <3242858943605340@naggum.no>
* Christopher C. Stacy
| LOOP was brought into Common Lisp from MACLISP, and I am sure
| there must be a LOOP library implementation that can be optionally
| loaded into GCL. LOOP hadn't made it into the Common Lisp standard,
| but it certainly existed in all its full-blown glory.
The interesting thing about GCL is its history. It dates back to Kyoto
Common Lisp (KCL), which was the first implementation of Common Lisp that
did not have access to the rich oral tradition of the Lisp community,
only the specification. This led to numerous interesting differences
from the expecations of those who did have access to the oral tradition,
and let to a much better specification the second time around with many
cleanup issues going back to the Kyoto implementation. The sad thing is
that this compiler never tracked the developments of the language and was
by and large abandoned. W. Schelter at the University of Austin picked
it up and made several enhancements to it and made it known as the Austin
Kyoto Common Lisp (AKCL), but as far as I know, he has left Common Lisp
and this world behind, and it was unsupported for several years before it
was resurrected by the FSF as GNU Common Lisp in preference to CMUCL for
political reasons.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> writes:
> KCL/AKCL [...] was unsupported for several years before it was
> resurrected by the FSF as GNU Common Lisp in preference to CMUCL
> for political reasons.
Could you provide some details about the political reasons? I'm just
curious because from what I perceive now (knowing CL for about two
years) CMUCL seems to be the better choice.
[I've bought two LispWorks licenses recently for various reasons and I
also very much respect AllegroCL although it's currently not quite in
my price range but I'm still amazed by the high quality of CMUCL,
especially if compared to other "free" software products. Also, for
certain kinds of applications it still seems to be the fastest CL
implementation to date.]
Thanks in advance,
Edi.
From: Erik Naggum
Subject: Re: Error: The function DEFCLASS is undefined
Date:
Message-ID: <3242904687940422@naggum.no>
* Edi Weitz
| Could you provide some details about the political reasons?
There is not much to say, really. The CMUCL people did not want to go
under the GPL, and those who wanted the GPL did not want CMUCL because it
would be nearly impossible to enforce the GPL for CMUCL.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> writes:
> that this compiler never tracked the developments of the language and was
> by and large abandoned. W. Schelter at the University of Austin picked
> it up and made several enhancements to it and made it known as the Austin
> Kyoto Common Lisp (AKCL), but as far as I know, he has left Common Lisp
> and this world behind, and it was unsupported for several years before it
> was resurrected by the FSF as GNU Common Lisp in preference to CMUCL for
> political reasons.
Bill Schelter was the official (RMS-approved) maintainer of GCL for
some time, as can be seen from his name on various GNU web pages - and
from the GCL references on his web pages. In fact, the archive at
ftp://ftp.ma.utexas.edu/pub/gcl still has GCL versions dating back to
1994
The FSF don't like CMUCL for being public domain[*] (they don't like
that people could fork non-'free' versions of it) but also they have a
desire that there be only one source for machine descriptions in GNU,
and that that be GCC - so they really want all their other compilers
to be able to compile to C, or plug into the GCC backend directly.
-dan
[*] Mostly public domain. Bits of it (LOOP, PCL) fall under the
MIT/X-consortium style of "almost public domain" licence
--
http://ww.telent.net/cliki/ - Link farm for free CL-on-Unix resources
From: Paul F. Dietz
Subject: Re: GCL (was Re: Error: The function DEFCLASS is undefined)
Date:
Message-ID: <3DA02C45.3020105@dls.net>
Daniel Barlow wrote:
> The FSF don't like CMUCL for being public domain[*] (they don't like
> that people could fork non-'free' versions of it) but also they have a
> desire that there be only one source for machine descriptions in GNU,
> and that that be GCC - so they really want all their other compilers
> to be able to compile to C, or plug into the GCC backend directly.
And because of this, you can get gcl up on, say, Power, ARM, or ia64,
while porting CMU CL to those architectures would be some amount
of work. Of course, CMU CL has the compensating benefit of
generating faster code.
The bit about going through GCC isn't quite correct -- there's a
dynamic code generation library called GNU Lightning that, as far as
I know, doesn't use GCC's machine descriptions.
Paul
"Paul F. Dietz" <·····@dls.net> writes:
> And because of this, you can get gcl up on, say, Power, ARM, or ia64,
> while porting CMU CL to those architectures would be some amount
> of work. Of course, CMU CL has the compensating benefit of
> generating faster code.
Yes, of course there are tradeoffs.
> The bit about going through GCC isn't quite correct -- there's a
> dynamic code generation library called GNU Lightning that, as far as
> I know, doesn't use GCC's machine descriptions.
The bit about going through GCC is based on a conversation I had (or,
more accurately, stood next to and didn't make any meaningful
contribution to) earlier this summer with RMS and Robert Strandh.
I didn't know about GNU Lightning - thanks for the pointer.
-dan
--
http://ww.telent.net/cliki/ - Link farm for free CL-on-Unix resources
Hi there.
> And because of this, you can get gcl up on, say, Power, ARM, or ia64,
And not forgetting the GCL ports to MS Windows and Windows CE, for which
platforms CMUCL has regrettably not made an appearance.
Cheers
Mike Thomas
"Mike Thomas" <······@brisbane.paradigmgeo.com> writes:
> And not forgetting the GCL ports to MS Windows and Windows CE, for which
> platforms CMUCL has regrettably not made an appearance.
As CMUCL already has a code generation backend for the x86, I doubt
that this is the reason it's not available for MS Windows. The
information at http://sbcl.sourceforge.net/ports.php suggests some
more likely explanations for no-windows-port. (Yes it's for SBCL, but
the two systems are fundamentally pretty similar at this level)
Windows CE is another matter, yes. There's a MIPS backend (admittedly
unsupported in present CMUCL versions, but has been resuscitated for
SBCL and "seems, with a little encouragement, to be able to build
itself"), but someone would have to write the necessary backends for
ARM/SH3/SH4/whatever else it runs on.
-dan
--
http://ww.telent.net/cliki/ - Link farm for free CL-on-Unix resources