Depends on your platform and financial situation. Given that you are
a newbie I assume you'll want a free compiler. The main ones seem to
be:
CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
OpenMCL http://openmcl.clozure.com/ OS X and Linux PPC
CMUCL http://www.cons.org/cmucl/ Linux, BSD
SBCL http://sbcl.sourceforge.net/ Linux, BSD
Hope that helps,
Justin Dubs
In article <····························@posting.google.com>,
······@eos.ncsu.edu (Justin Dubs) wrote:
> Depends on your platform and financial situation. Given that you are
> a newbie I assume you'll want a free compiler. The main ones seem to
> be:
>
> CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
CLISP is basically broken on Mac OS X. Certainly nothing
for a newbie on Mac OS X.
> OpenMCL http://openmcl.clozure.com/ OS X and Linux PPC
> CMUCL http://www.cons.org/cmucl/ Linux, BSD
> SBCL http://sbcl.sourceforge.net/ Linux, BSD
SBCL does run under Mac OS X.
Rainer Joswig <······@lispmachine.de> wrote in message news:<····························@news.fu-berlin.de>...
> In article <····························@posting.google.com>,
> ······@eos.ncsu.edu (Justin Dubs) wrote:
>
> > Depends on your platform and financial situation. Given that you are
> > a newbie I assume you'll want a free compiler. The main ones seem to
> > be:
> >
> > CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
>
> CLISP is basically broken on Mac OS X. Certainly nothing
> for a newbie on Mac OS X.
Why do you believe this? CLISP 2.29 compiles out of the box on OS X
and seems to work fine. What specific problems are you having?
>
> > OpenMCL http://openmcl.clozure.com/ OS X and Linux PPC
> > CMUCL http://www.cons.org/cmucl/ Linux, BSD
> > SBCL http://sbcl.sourceforge.net/ Linux, BSD
>
> SBCL does run under Mac OS X.
You're right. My mistake. I have 0.8.2.7 on my PowerBook right now.
Justin Dubs
From: Christophe Rhodes
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <sqad98cuzz.fsf@lambda.jcn.srcf.net>
······@eos.ncsu.edu (Justin Dubs) writes:
> Rainer Joswig <······@lispmachine.de> wrote in message news:<····························@news.fu-berlin.de>...
>> SBCL does run under Mac OS X.
>
> You're right. My mistake. I have 0.8.2.7 on my PowerBook right now.
One of the good things about exposing one's work to the wider world is
that it gets exposed to new situations... or, to put it less
obliquely, sbcl-0.8.2.7 is good enough for compiling a later version,
but testing revealed several instances where the ppc backend was less
than optimal. I'd recommend an upgrade (and if you do, you get the
extra bonus of being able to feed back bug reports to the developers
should you find anything newly wrong). As Rainer says, compiling sbcl
is meant to be absolutely trivial; there should be nothing to
fear. :-)
Cheers,
Christophe
--
http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)
In article <····························@posting.google.com>,
······@eos.ncsu.edu (Justin Dubs) wrote:
> Rainer Joswig <······@lispmachine.de> wrote in message news:<····························@news.fu-berlin.de>...
> > In article <····························@posting.google.com>,
> > ······@eos.ncsu.edu (Justin Dubs) wrote:
> >
> > > Depends on your platform and financial situation. Given that you are
> > > a newbie I assume you'll want a free compiler. The main ones seem to
> > > be:
> > >
> > > CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
> >
> > CLISP is basically broken on Mac OS X. Certainly nothing
> > for a newbie on Mac OS X.
>
> Why do you believe this? CLISP 2.29 compiles out of the box on OS X
> and seems to work fine. What specific problems are you having?
CLisp 2.29 is a bit old. On my machines I could compiled none
of the releases of CLisp. Not one. Even the binary of CLisp 2.31
on sourceforge does not run - the installation simply does not work.
I've read the discussions on the Clisp mailing list how people
seem to manage somehow to compile a newer CLisp. I can't reproduce
that. I tried it on two machines running Mac OS X 10.2.6 and
the latest developer stuff.
Actually I've given up on CLisp until it is a really supported
port on Mac OS X. Not that I need it - there are plenty
other Lisps on that platform now. I guess eventually some
version of CLisp will be easier to install...
In contrast I've managed to compile OpenMCL and I also
managed to compile SBCL (which is actually REALLY easy).
That something complicated like SBCL compiles THAT
simple is really good work from the maintainers!!!
> > > OpenMCL http://openmcl.clozure.com/ OS X and Linux PPC
> > > CMUCL http://www.cons.org/cmucl/ Linux, BSD
> > > SBCL http://sbcl.sourceforge.net/ Linux, BSD
> >
> > SBCL does run under Mac OS X.
>
> You're right. My mistake. I have 0.8.2.7 on my PowerBook right now.
I'm using 0.8.3.38 .
Rainer Joswig
From: Douglas Philips
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <3F637DD4.3040904@mac.com>
Rainer Joswig wrote:
> In article <····························@posting.google.com>,
> ······@eos.ncsu.edu (Justin Dubs) wrote:
>>Depends on your platform and financial situation. Given that you are
>>a newbie I assume you'll want a free compiler. The main ones seem to
>>be:
>>
>>CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
>
>
> CLISP is basically broken on Mac OS X. Certainly nothing
> for a newbie on Mac OS X.
Huh? 2.31 builds on Mac OS/X and seems to be just fine.
In what way is it broken? And why is CLisp not suitable for a newbie (on
OS/X or anywhere else?)
<D\'gou
In article <················@mac.com>, Douglas Philips <····@mac.com>
wrote:
> Rainer Joswig wrote:
> > In article <····························@posting.google.com>,
> > ······@eos.ncsu.edu (Justin Dubs) wrote:
> >>Depends on your platform and financial situation. Given that you are
> >>a newbie I assume you'll want a free compiler. The main ones seem to
> >>be:
> >>
> >>CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
> >
> >
> > CLISP is basically broken on Mac OS X. Certainly nothing
> > for a newbie on Mac OS X.
>
> Huh? 2.31 builds on Mac OS/X and seems to be just fine.
Since when? Which version?
Additionally I just downloaded the binary and it
just CANNOT work. No way. The makefile is broken.
Error in pathnames. Lib that is not installed by default on
Mac OS X is referenced. and so on.
Compare that whole installation procedure with
SBCL where you just untar it and it just works.
http://sourceforge.net/project/showfiles.php?group_id=1373&release_id=180115
You can recompile SBCL with just:
get newest sources from CVS
sudo gcc_select 3
limit stacksize 8192
cd sbcl
./make.sh
And then you have recompiled the whole stuff.
> In what way is it broken? And why is CLisp not suitable for a newbie (on
> OS/X or anywhere else?)
Installation of 2.31 is non-trivial. Sure its just me. Others seem
to be struggling too...
I read http://article.gmane.org/gmane.lisp.clisp.devel/10389
and your answer.
Or earlier this article:
http://article.gmane.org/gmane.lisp.clisp.general/7436
Or the announcement of 2.31 which does not even mention Mac OS X as
a platform GNU CLISP runs on:
http://article.gmane.org/gmane.lisp.clisp.announce/22
From the latest unix/PLATFORMS file:
---
On Apple PowerPC running MacOS X Server:
You can get C development tools from either
Fink - < http://fink.sourceforge.net > (Free Software ported to Darwin)
or Apple - < http://developer.apple.com/tools >.
The standard "/bin/sh" shell has at least two bugs which make it
unusable for the configuration scripts. As a workaround, you have to
set the environment variable CONFIG_SHELL to "/bin/bash", and start
"$CONFIG_SHELL ./configure ..." instead of "./configure ...".
The default stack size limit is 512 KB, which is too small for bootstrapping
CLISP. Even 1 MB is too small. Try "ulimit -S -s 8192" before starting "make"
or "limit stacksize 8192" is you are using a CSH-derivative.
Remove all optimization options ("-O", "-O2") from the CC and CFLAGS variables
in the Makefile. Apple's cc crashes when compiling eval.d with optimization.
FFI does not work yet, you will have to ignore the "avcall-*.h" errors
in ./configure or run './configure --without-dynamic-ffi'.
Dec2002 Developer tools with the gcc 3.3 update work OOTB, but
older versions of GCC might require the `--traditional-cpp' option.
If you get weird compilation errors (like #else mismatch), try it.
---
I would hope that this means also that it would run under Mac OS X client.
These instructions are definitely NOT looking like NEWBIE stuff.
Maybe we have different definitions of 'newbie'. ? ;-)
From: Douglas Philips
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <3F638AAA.1080304@mac.com>
Rainer Joswig wrote:
> In article <················@mac.com>, Douglas Philips <····@mac.com>
> wrote:
>>Huh? 2.31 builds on Mac OS/X and seems to be just fine.
>
> Since when? Which version?
Since before 2.31 was released (the anon cvs head from sf.net).
10.2.6 with the latest apple development tools.
You can build CLISP with "just":
get newest sources from CVS
(get latest dev. tools from Apple if you don't have them).
(get readline gnu library if you care, I had it already for other stuff).
limit stack 8192
setenv CONFIG_SHELL /bin/bash
$CONFIG_SHELL ./configure --without-dynamic-ffi build-directory-to-use
and follow the instructions that the configure script prints.
Once you are set up to (the first two/three steps), its not hard.
Whether or not you consider installing the apple tools hard, that nots a
CLISP issue.
> Additionally I just downloaded the binary and it
> just CANNOT work. No way. The makefile is broken.
> Error in pathnames. Lib that is not installed by default on
> Mac OS X is referenced. and so on.
Interesting. Its highly likely that I produced that binary, so I would
really like to know in what way it doesn't work for you so that I can
try to fix it. (We can take this offlist).
> Installation of 2.31 is non-trivial. Sure its just me. Others seem
> to be struggling too...
>
> I read http://article.gmane.org/gmane.lisp.clisp.devel/10389
> and your answer.
You seem to confuse installing CLISP with building it. Building it isn't
hard, but that's not the same as just installing it.
> Or the announcement of 2.31 which does not even mention Mac OS X as
> a platform GNU CLISP runs on:
> http://article.gmane.org/gmane.lisp.clisp.announce/22
Yeah, that bugged me a bit, but without FFI support (I don't have the
interest, nor if I did, the time, to do it myself), maybe that's reasonable.
> I would hope that this means also that it would run under Mac OS X client.
> These instructions are definitely NOT looking like NEWBIE stuff.
> Maybe we have different definitions of 'newbie'. ? ;-)
Hmmm, well, I guess I could claim how painfully hard it was to find Mac
OS X references on the SBCL site, but what would that prove? I never did
(this was a few weeks ago).
I'd rather just figure out how to make the installable binaries work on
more systems.
<D\'gou
In article <················@mac.com>, Douglas Philips <····@mac.com>
wrote:
...snip...
> > Additionally I just downloaded the binary and it
> > just CANNOT work. No way. The makefile is broken.
> > Error in pathnames. Lib that is not installed by default on
> > Mac OS X is referenced. and so on.
>
> Interesting. Its highly likely that I produced that binary, so I would
> really like to know in what way it doesn't work for you so that I can
> try to fix it. (We can take this offlist).
[RJPBG4:/Lisp/clisp] joswig% ls -l
total 6216
-rw-r--r-- 1 joswig staff 3180601 Sep 10 21:13 clisp-2.31-ppc-powerpc-darwin-6.6.tar.gz
[RJPBG4:/Lisp/clisp] joswig% gunzip clisp-2.31-ppc-powerpc-darwin-6.6.tar.gz
[RJPBG4:/Lisp/clisp] joswig% ls -l
total 23560
-rw-r--r-- 1 joswig staff 12062720 Sep 10 21:13 clisp-2.31-ppc-powerpc-darwin-6.6.tar
[RJPBG4:/Lisp/clisp] joswig% tar xf clisp-2.31-ppc-powerpc-darwin-6.6.tar
[RJPBG4:/Lisp/clisp] joswig% ls
clisp-2.31 clisp-2.31-ppc-powerpc-darwin-6.6.tar
[RJPBG4:/Lisp/clisp] joswig% cd clisp-2.31
[RJPBG4:/Lisp/clisp/clisp-2.31] joswig% ls
ANNOUNCE GNU-GPL Makefile README README.es base data emacs linkkit src
COPYRIGHT MAGIC.add NEWS README.de SUMMARY clisp-link doc full regexp syscalls
[RJPBG4:/Lisp/clisp/clisp-2.31] joswig% make
cc -O base/modules.o base/lisp.a base/libcharset.a -lreadline -lncurses base//usr/local/lib/libiconv.dylib -o base/lisp.run
cc: base//usr/local/lib/libiconv.dylib: No such file or directory
make: *** [base/lisp.run] Error 1
[RJPBG4:/Lisp/clisp/clisp-2.31] joswig%
hmm...
> > Installation of 2.31 is non-trivial. Sure its just me. Others seem
> > to be struggling too...
> >
> > I read http://article.gmane.org/gmane.lisp.clisp.devel/10389
> > and your answer.
>
> You seem to confuse installing CLISP with building it. Building it isn't
> hard, but that's not the same as just installing it.
Well, since I could not install from binaries (or at the time when
I first tried, there were no binaries), I tried to build it.
...snip...
From: Douglas Philips
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <3F6391F5.1060506@mac.com>
[RJPBG4:/Lisp/clisp] joswig% cd clisp-2.31
> [RJPBG4:/Lisp/clisp/clisp-2.31] joswig% ls
> ANNOUNCE GNU-GPL Makefile README README.es base data emacs linkkit src
> COPYRIGHT MAGIC.add NEWS README.de SUMMARY clisp-link doc full regexp syscalls
> [RJPBG4:/Lisp/clisp/clisp-2.31] joswig% make
> cc -O base/modules.o base/lisp.a base/libcharset.a -lreadline -lncurses base//usr/local/lib/libiconv.dylib -o base/lisp.run
> cc: base//usr/local/lib/libiconv.dylib: No such file or directory
> make: *** [base/lisp.run] Error 1
> [RJPBG4:/Lisp/clisp/clisp-2.31] joswig%
Interesting. If you change the makefile and replace the
base//usr/local/lib/libiconv.dylib (and the
full//usr/local/lib/libiconv.dylib) to -liconv
does everything else work? (I got warnings, but it worked). I'll try to
have a new tar file with this fix for you by tomorrow, but it would be
very helpful to know if this is the "only" problem).
Thanks for your help.
I'm not officially part of the CLISP team, but I really like it, and I
really like Mac OS/X, so if I can help out, please let me know and I'll
do what I can.
<D\'gou
In article <················@mac.com>, Douglas Philips <····@mac.com>
wrote:
> [RJPBG4:/Lisp/clisp] joswig% cd clisp-2.31
> > [RJPBG4:/Lisp/clisp/clisp-2.31] joswig% ls
> > ANNOUNCE GNU-GPL Makefile README README.es base data emacs linkkit src
> > COPYRIGHT MAGIC.add NEWS README.de SUMMARY clisp-link doc full regexp syscalls
> > [RJPBG4:/Lisp/clisp/clisp-2.31] joswig% make
> > cc -O base/modules.o base/lisp.a base/libcharset.a -lreadline -lncurses base//usr/local/lib/libiconv.dylib -o base/lisp.run
> > cc: base//usr/local/lib/libiconv.dylib: No such file or directory
> > make: *** [base/lisp.run] Error 1
> > [RJPBG4:/Lisp/clisp/clisp-2.31] joswig%
>
> Interesting. If you change the makefile and replace the
> base//usr/local/lib/libiconv.dylib (and the
> full//usr/local/lib/libiconv.dylib) to -liconv
> does everything else work? (I got warnings, but it worked). I'll try to
> have a new tar file with this fix for you by tomorrow, but it would be
> very helpful to know if this is the "only" problem).
>
> Thanks for your help.
> I'm not officially part of the CLISP team, but I really like it, and I
> really like Mac OS/X, so if I can help out, please let me know and I'll
> do what I can.
>
> <D\'gou
>
>
changed and then:
[RJPBG4:/Lisp/clisp/clisp-2.31] joswig% make
cc -O base/modules.o base/lisp.a base/libcharset.a -lncurses -liconv -o base/lisp.run
ld: can't locate file for: -liconv
Rainer Joswig <······@lispmachine.de> wrote in message news:<····························@news.fu-berlin.de>...
> CLISP is basically broken on Mac OS X. Certainly nothing
> for a newbie on Mac OS X.
Let me second Rainer's negative assesment of building CLISP on Mac OS
X 10.2.6 with the latest developer tools. I get precisely the same
errors that Rainer reports subsequently in this thread.
If it is possible to build on Mac OS X, one would think that the
install instructions should detail the various non-obvious things one
must do in order for the build to succeed, rather than leaving the
user with the impression that it is just broken.
This is thrown into sharp contrast by the ease of building sbcl under
Mac OS X.
From: Greg Menke
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <m3d6e3azmo.fsf@europa.pienet>
·······@mediaone.net (Raffael Cavallaro) writes:
> Rainer Joswig <······@lispmachine.de> wrote in message news:<····························@news.fu-berlin.de>...
>
>
> > CLISP is basically broken on Mac OS X. Certainly nothing
> > for a newbie on Mac OS X.
>
> Let me second Rainer's negative assesment of building CLISP on Mac OS
> X 10.2.6 with the latest developer tools. I get precisely the same
> errors that Rainer reports subsequently in this thread.
>
> If it is possible to build on Mac OS X, one would think that the
> install instructions should detail the various non-obvious things one
> must do in order for the build to succeed, rather than leaving the
> user with the impression that it is just broken.
>
> This is thrown into sharp contrast by the ease of building sbcl under
> Mac OS X.
I've built and installed CLISP on a number of boxes, x86 Linux, PPC
Linux, Solaris, etc.. and have had no problems EXCEPT for one box
which absolutely, positively would not run CLISP. No other software
had trouble on it (Debian 2.2 kernel on a Cyrix 200)- even changing
kernels & c libraries was of no help. CLISP would build fine, but
always segfaulted at startup, regardless of the ./configure options-
or even if I used the Debian CLISP package.
So I guess my point is, CLISP may be suffering from a few subtle
pathological cases which are hard to find because its difficult to
provoke them.
Gregm
From: Douglas Philips
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <3F6526FB.70801@mac.com>
Greg Menke wrote:
> ·······@mediaone.net (Raffael Cavallaro) writes:
>>If it is possible to build on Mac OS X, one would think that the
>>install instructions should detail the various non-obvious things one
>>must do in order for the build to succeed, rather than leaving the
>>user with the impression that it is just broken.
I'm confused as to whether you are referring to really building CLISP
from scratch, or if you mean the binary installation process (which
involves doing make and a few other 'build' like things).
>>
>>This is thrown into sharp contrast by the ease of building sbcl under
>>Mac OS X.
"Inconceivable!"
"You keep using that word. I do not think it means what you think it means."
What Mac OS X support. Nothing mentioned on any of: http://sbcl.sf.net
http://sbcl.sourceforge.net/ports.php (which seems quite old) nd
http://sbcl-internals.cliki.net/PPC
says:
What about MacOS X
There is some interest in a OSX port. It should mostly not be rocket
science: I (Daniel Barlow) can give pointers (ask on sbcl-devel?) but
I'm unlikely to actually write it myself unless someone sends me (a)
financial incentive, or (b) a sufficiently well-endowed tiBook
MacOS 10.1 apparently has a problem with the SA_SIGINFO flag for signal
handlers in that it either doesn't support it or doesnt do something
useful if it does. This would be a problem for SBCL, which needs to be
able to read and change register contents in the sigcontext for e.g. GC.
10.2 is supposed to fix this; alternatively the OpenMCL solution (using
another Mach task to frob the stack and fake a sigcontext on it) could
be borrowed.
Ye Gods, why would I even consider SBCL over CLISP? At least I can get
CLisp to build, at least I can get an old (2.29) version binaries.
If one presses on anyways, it seems that:
http://sourceforge.net/project/showfiles.php?group_id=1373
does list a binary, which I'm downloading now.
Look, I'm not saying that CLisp is perfect, but its rather ironic the
invective lobbed at it esp. by comparison with SBCL when SBCL looks to
what a newbie may find (see the pages listed above) as dead, and
certainly not close for MacOS/X.
I'm currently working to resolve the 2.31 binary install issue, but time
is...
> I've built and installed CLISP on a number of boxes, x86 Linux, PPC
> Linux, Solaris, etc.. and have had no problems EXCEPT for one box
> which absolutely, positively would not run CLISP. No other software
> had trouble on it (Debian 2.2 kernel on a Cyrix 200)- even changing
> kernels & c libraries was of no help. CLISP would build fine, but
> always segfaulted at startup, regardless of the ./configure options-
> or even if I used the Debian CLISP package.
>
> So I guess my point is, CLISP may be suffering from a few subtle
> pathological cases which are hard to find because its difficult to
> provoke them.
At least I'm not the only one to have success on multiple platforms
(including MacOS/X). Thanks Greg for the data point!
<D\'gou
From: Christophe Rhodes
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <sq3ceyxx0j.fsf@lambda.jcn.srcf.net>
Douglas Philips <····@mac.com> writes:
> What Mac OS X support. Nothing mentioned on any of: http://sbcl.sf.net
> http://sbcl.sourceforge.net/ports.php (which seems quite old) nd
> http://sbcl-internals.cliki.net/PPC
> says:
> What about MacOS X
Thanks. I've updated these pages to mention Darwin support.
> Ye Gods, why would I even consider SBCL over CLISP? At least I can get
> CLisp to build, at least I can get an old (2.29) version binaries.
I'd hope that one would consider at least looking at both. Sorry that
the perception you gained from the various websites you looked at was
that MacOS X was an unsupported platform.
> Look, I'm not saying that CLisp is perfect, but its rather ironic the
> invective lobbed at it esp. by comparison with SBCL when SBCL looks to
> what a newbie may find (see the pages listed above) as dead, and
> certainly not close for MacOS/X.
I hope that the current situation is better, but if you do find
another "newbie" route into the sbcl information web which gives the
impression that the project is dead or that platforms which should be
mentioned aren't, feedback to the developer lists is always
appreciated.
Christophe
--
http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)
From: Douglas Philips
Subject: Re: hi all i am new to this so what is a good compiler to use forlisp
Date:
Message-ID: <3F65C196.5000508@mac.com>
Christophe Rhodes wrote:
>>What Mac OS X support. Nothing mentioned on any of: http://sbcl.sf.net
>>http://sbcl.sourceforge.net/ports.php (which seems quite old) nd
>>http://sbcl-internals.cliki.net/PPC
>
> Thanks. I've updated these pages to mention Darwin support.
Thanks.
> I hope that the current situation is better, but if you do find
> another "newbie" route into the sbcl information web which gives the
> impression that the project is dead or that platforms which should be
> mentioned aren't, feedback to the developer lists is always
> appreciated.
Sure, though mentioning "Darwin" is not necessarily going to tell a
newbie that OS/X is supported. My impression is that the Darwin / OS/X
link is mostly something developers are aware of, but ymmv.
<D\'gou
From: Daniel Barlow
Subject: Re: hi all i am new to this so what is a good compiler to use forlisp
Date:
Message-ID: <87u1775y5u.fsf@noetbook.telent.net>
Douglas Philips <····@mac.com> writes:
> Christophe Rhodes wrote:
>> Thanks. I've updated these pages to mention Darwin support.
[...]
> Sure, though mentioning "Darwin" is not necessarily going to tell a
> newbie that OS/X is supported. My impression is that the Darwin / OS/X
> link is mostly something developers are aware of, but ymmv.
I have to wonder why you need a compiler unless you're a developer.
How many Mac users know their system is running "OS/X" anyway?
Perhaps that's still too technical, and we should be saying that SBCL
runs on "many translucent computers, and also some that are made out
of titanium".
Your opinion is valued as always.
-dan
--
http://www.cliki.net/ - Link farm for free CL-on-Unix resources
From: Douglas Philips
Subject: Re: hi all i am new to this so what is a good compiler to useforlisp
Date:
Message-ID: <3F6C6CC7.8070803@mac.com>
Daniel Barlow wrote:
> I have to wonder why you need a compiler unless you're a developer.
????? Hopefully you don't. What led to this question?
> How many Mac users know their system is running "OS/X" anyway?
That's how Apple brands it. Much more so that "Darwin".
<D\'gou
In article <················@mac.com>, Douglas Philips <····@mac.com>
wrote:
> Daniel Barlow wrote:
> > I have to wonder why you need a compiler unless you're a developer.
>
> ????? Hopefully you don't. What led to this question?
The simple fact that anyone who has ever popped open a terminal under OS
X has seen the words "Welcome to Darwin" at least once before, and
anyone who is a developer has probably opened a terminal at least once
before asking questions like this.
Bottom line is - if you've never opened a terminal, then you probably
aren't looking for a CL implementation.
>
> > How many Mac users know their system is running "OS/X" anyway?
>
> That's how Apple brands it. Much more so that "Darwin".
Wrong. Apple brands it as "Mac OS X". No slash. You will and never have
seen the slash present on Apple's web site nor on any of the competent
news sites. Anyone who continues to spell it this way is obviously
unwilling to adjust his head to match reality anyway.
And as I mentioned above, anyone who has opened a terminal at least once
is aware of the fact that Apple does brand the underlying BSD system
"Darwin".
--
Brian Mastenbrook
http://cs.indiana.edu/~bmastenb/
Brian Mastenbrook <··············@indiana.edu> writes:
> In article <················@mac.com>, Douglas Philips <····@mac.com>
> wrote:
>
> > Daniel Barlow wrote:
> > > I have to wonder why you need a compiler unless you're a developer.
> >
> > ????? Hopefully you don't. What led to this question?
>
> The simple fact that anyone who has ever popped open a terminal under OS
> X has seen the words "Welcome to Darwin" at least once before, and
> anyone who is a developer has probably opened a terminal at least once
> before asking questions like this.
>
> Bottom line is - if you've never opened a terminal, then you probably
> aren't looking for a CL implementation.
What a wacky conclusion! If you use the system in the suggested
manner, you can write kernel extensions, even, without touching a
terminal. Even once.
I think a much better point would be that Apple heavily pushes the
whole "Darwin" thing in its promo material for developers, even if
normal users might never see the term.
--
/|_ .-----------------------.
,' .\ / | No to Imperialist war |
,--' _,' | Wage class war! |
/ / `-----------------------'
( -. |
| ) |
(`-. '--.)
`. )----'
From: ······@nordebo.com
Subject: Re: hi all i am new to this so what is a good compiler to useforlisp
Date:
Message-ID: <m21xub5m60.fsf@patrik.nordebo.com>
···@famine.OCF.Berkeley.EDU (Thomas F. Burdick) writes:
> Brian Mastenbrook <··············@indiana.edu> writes:
> > Bottom line is - if you've never opened a terminal, then you probably
> > aren't looking for a CL implementation.
>
> What a wacky conclusion! If you use the system in the suggested
> manner, you can write kernel extensions, even, without touching a
> terminal. Even once.
Well, you might write kernel extensions all in the comfy OS X
environment, but you're going to need a terminal to run SBCL (as there
is no fancy packaging). Of course you could also use it in Emacs or
such, but, err, let's say that's beside the point. :-)
Anyway, I think describing the current state of the port as a Darwin
port rather than an OS X port is pretty fair, as there is absolutely
no integration whatsoever with anything outside what Apple calls
Darwin. Thus far SBCL treats OS X just like any other Unix, and that
is what Darwin is to me - the Unix substrate which underlies the OS X
experience. At some point we will hopefully have support for the other
fancy stuff (Cocoa and whatnot), and we might talk about supporting OS
X, but currently saying it does is a bit misleading, IMHO.
--
Patrik Nordebo
From: Douglas Philips
Subject: Re: SBCL on Darwin (was; Re: hi all i am new to this so what is a good compiler to useforlisp
Date:
Message-ID: <3F6CC5EE.4050201@mac.com>
······@nordebo.com wrote:
> Of course you could also use it in Emacs or
> such, but, err, let's say that's beside the point. :-)
Just as you can run TeX from TeXShop and never see or care about a
command line. Really, the command line is a convenience, but since when
was the almighty command line of any interest to Lispers, living in
their own cozy isolated Lisp Machine if I could afford it, but Lisp IDE
of some kind anyways, world? (And no, I'm not asking this of Patrik
directly).
> Anyway, I think describing the current state of the port as a Darwin
> port rather than an OS X port is pretty fair, as there is absolutely
> no integration whatsoever with anything outside what Apple calls
> Darwin.
Anyway, that does seem a fair characterization. All depends on how you
want to be searched, er, or rather, found.
In any event, I see Apple marketting OS X Server machines, not Darwin
Server machines.
>At some point we will hopefully have support for the other
> fancy stuff (Cocoa and whatnot), and we might talk about supporting OS
> X, but currently saying it does is a bit misleading, IMHO.
Can SBCL link with native OS libraries?
<D\'gou
From: ······@nordebo.com
Subject: Re: SBCL on Darwin (was; Re: hi all i am new to this so what is a good compiler to useforlisp
Date:
Message-ID: <m2vfrn3zdb.fsf@patrik.nordebo.com>
Douglas Philips <····@mac.com> writes:
> > Anyway, I think describing the current state of the port as a Darwin
> > port rather than an OS X port is pretty fair, as there is absolutely
> > no integration whatsoever with anything outside what Apple calls
> > Darwin.
>
> Anyway, that does seem a fair characterization. All depends on how you
> want to be searched, er, or rather, found.
Yes, well, probably it is good to mention that the Darwin port does in
fact run on OS X, as is currently the case on the SBCL front page on
Sourceforge.
> >At some point we will hopefully have support for the other
> > fancy stuff (Cocoa and whatnot), and we might talk about supporting OS
> > X, but currently saying it does is a bit misleading, IMHO.
>
> Can SBCL link with native OS libraries?
Yes, this is all "user-level" stuff, so to speak, it shouldn't take
any hacking of SBCL itself (modulo bugs, of course). But it is
probably quite a bit of work to get there.
--
Patrik Nordebo
From: Pascal Bourguignon
Subject: Re: SBCL on Darwin (was; Re: hi all i am new to this so what is a good compiler to useforlisp
Date:
Message-ID: <87oexeut2t.fsf@thalassa.informatimago.com>
Douglas Philips <····@mac.com> writes:
> > Anyway, I think describing the current state of the port as a Darwin
> > port rather than an OS X port is pretty fair, as there is absolutely
> > no integration whatsoever with anything outside what Apple calls
> > Darwin.
>
> Anyway, that does seem a fair characterization. All depends on how you
> want to be searched, er, or rather, found.
>
> In any event, I see Apple marketting OS X Server machines, not Darwin
> Server machines.
And rightly so, since what they sell is actually GUI administration
tools that <hype>allows a simple user to administrate the server, or
that allow a unix administrator to administration 100x more servers
than of the concurence</hype>. Note that Darwin is not sold by Apple:
it's free! It's used by Apple as a base for their MacOSX, that's all.
--
__Pascal_Bourguignon__
http://www.informatimago.com/
Do not adjust your mind, there is a fault in reality.
In article <··············@patrik.nordebo.com>, ······@nordebo.com
wrote:
> ···@famine.OCF.Berkeley.EDU (Thomas F. Burdick) writes:
>
> > Brian Mastenbrook <··············@indiana.edu> writes:
> > > Bottom line is - if you've never opened a terminal, then you probably
> > > aren't looking for a CL implementation.
> >
> > What a wacky conclusion! If you use the system in the suggested
> > manner, you can write kernel extensions, even, without touching a
> > terminal. Even once.
>
> Well, you might write kernel extensions all in the comfy OS X
> environment, but you're going to need a terminal to run SBCL (as there
> is no fancy packaging). Of course you could also use it in Emacs or
> such, but, err, let's say that's beside the point. :-)
>
> Anyway, I think describing the current state of the port as a Darwin
> port rather than an OS X port is pretty fair, as there is absolutely
> no integration whatsoever with anything outside what Apple calls
> Darwin. Thus far SBCL treats OS X just like any other Unix, and that
> is what Darwin is to me - the Unix substrate which underlies the OS X
> experience. At some point we will hopefully have support for the other
> fancy stuff (Cocoa and whatnot), and we might talk about supporting OS
> X, but currently saying it does is a bit misleading, IMHO.
Apple's Darwin is also a real operating system that runs
on PowerPC and x86 (!!). The x86 version is available.
I personally don't know anybody who actually runs Darwin
on a PC, but it is possible - given that the hardware
is supported.
On top of Darwin there is a lot of additional stuff that
makes Mac OS X. Although it is said that much of Mac OS X
actually runs on Intel machines, no version has ever been
published.
Links:
http://developer.apple.com/darwin/
See also http://www.opendarwin.org/
Apple has published the source here:
http://www.opensource.apple.com/darwinsource/index.html
a x86 version of Darwin is here: http://darwin.monkeyvoodoo.net/
See also the GNU-Darwin project: http://gnu-darwin.sourceforge.net/
From: Douglas Philips
Subject: Re: hi all i am new to this so what is a good compiler to useforlisp
Date:
Message-ID: <3F6CC3B2.50405@mac.com>
Brian Mastenbrook wrote:
> Wrong. Apple brands it as "Mac OS X". No slash. You will and never have
> seen the slash present on Apple's web site nor on any of the competent
> news sites. Anyone who continues to spell it this way is obviously
> unwilling to adjust his head to match reality anyway.
Ohhhhh, good one there, nothing like a gratuitous dose of ad hominem
when you're not sure of your technical merit, including a dollup of
stoopid spelling nit-picking. Whether spelt with or without a slash, OS
X is what you see everywhere.
But then if SBCL is only for "developers" who cares.
<D\'gou
From: Christophe Rhodes
Subject: Re: hi all i am new to this so what is a good compiler to useforlisp
Date:
Message-ID: <sqzngymmun.fsf@lambda.jcn.srcf.net>
Douglas Philips <····@mac.com> writes:
> But then if SBCL is only for "developers" who cares.
SBCL is explicitly not only for developers of SBCL. It's meant to be
a living implementation of Common Lisp that people do things with.
On the other hand, it is an implementation of Common Lisp, and
probably the thing that most people do with Common Lisp these days is
develop applications for it. So in that sense it is at least
primarily for developers.
At the moment I'm having difficulty in finding a usage case for a
Common Lisp that doesn't at any point involve having a developer in
the picture. Could you explain your concern -- particularly, could
you give an example of someone who specifically wants Common Lisp but
isn't a developer? If this is a sufficiently general scenario, it
would certainly be worth my while creating a page with enough
background information for this user to divine the information that he
needs.
Thanks,
Christophe
--
http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)
From: Douglas Philips
Subject: Re: hi all i am new to this so what is a good compiler touseforlisp
Date:
Message-ID: <3F6EF43A.90203@mac.com>
Christophe Rhodes wrote:
> At the moment I'm having difficulty in finding a usage case for a
> Common Lisp that doesn't at any point involve having a developer in
> the picture. Could you explain your concern -- particularly, could
> you give an example of someone who specifically wants Common Lisp but
> isn't a developer? If this is a sufficiently general scenario, it
> would certainly be worth my while creating a page with enough
> background information for this user to divine the information that he
> needs.
Unlike 'gcc' which comes with (whether or not installed automaticly by
your favorite distro of) Linux, someone looking to use/run a
written-in-CL package/solution/etc. needs to find a CL to use...
(on OS X this is also true since prebuilt Mac OS X binaries still aren't
as popular, but the Apple Web site provides free developer tools from
which you can get gcc, but not (IIRC) CL).
Given that SBCL is still versioned "less than 1" it might not matter as
much either. What that "really means" is less important than what it
will appear to mean to someone looking for a CL to use (rather than to
develop the CL system itself).
<D\'gou
Douglas Philips writes:
> Unlike 'gcc' which comes with (whether or not installed automaticly by
> your favorite distro of) Linux, someone looking to use/run a
> written-in-CL package/solution/etc. needs to find a CL to use...
Not necessarily. Most, if not all, the free Common Lisp
implementations, at least under Linux, provide ways for *easily*
packaging an application that can be run on a completely different
computer without further support files.
I have done that just the other day with CMUCL 18e. A good friend of
mine wants me to write some Lisp software for him (and, by the way, he
*insists* on paying me :) Since I told him that I will probably use
CLIM, I put together a McCLIM demo in a matter of minutes.
I dumped a CMUCL image with McCLIM and a few of its demos, copied the
`lisp' executable, created a shell one liner, put all of this into a
tarball, burnt a CD and run the whole thing on my friend's PC.
The latest CVS versions of both SBCL and CMUCL even provide a Lisp
form for automating the process and generating a "standalone"
application consisting of just one file. Folks, shipping applications
with most free Common Lisp implementations--let's not forget CLISP--is
*straightforward*.
Paolo
P.S.
Before I forget: perpetual kudos to the McCLIM developers.
--
Paolo Amoroso <·······@mclink.it>
From: Douglas Philips
Subject: Re: hi all i am new to this so what is a good compiler touseforlisp
Date:
Message-ID: <3f6f5be0$1_3@news.nauticom.net>
Paolo Amoroso wrote:
> Not necessarily. Most, if not all, the free Common Lisp
> implementations, at least under Linux, provide ways for *easily*
> packaging an application that can be run on a completely different
> computer without further support files.
I was making a nod to gcc/Linux by way of comparison with CL and Mac OS
X. AFAIK, CMUCL doesn't run on Darwin/OS X and SBCL was just ported. I
would be surprised if it could already dump native executables for OS X.
Then again, maybe I wouldn't be surprised.
<D\'gou
From: ······@nordebo.com
Subject: Re: hi all i am new to this so what is a good compiler touseforlisp
Date:
Message-ID: <m2k77y3r6x.fsf@patrik.nordebo.com>
Douglas Philips <····@mac.com> writes:
> I was making a nod to gcc/Linux by way of comparison with CL and Mac
> OS X. AFAIK, CMUCL doesn't run on Darwin/OS X and SBCL was just
> ported. I would be surprised if it could already dump native
> executables for OS X.
SBCL can't dump "native executables" (yet, anyway), but what Paolo
described was not dumping of a "native executable", but rather
generating a core (image) file containing the app, bundling the
runtime executable and running things from a shell script. This can be
done equally easily with SBCL. Packaging it up as a .app style
application should also be fairly easy, though I've never written an
app like that myself so I don't know what might be involved.
--
Patrik Nordebo
In article <··············@patrik.nordebo.com>, <······@nordebo.com>
wrote:
> Douglas Philips <····@mac.com> writes:
> > I was making a nod to gcc/Linux by way of comparison with CL and Mac
> > OS X. AFAIK, CMUCL doesn't run on Darwin/OS X and SBCL was just
> > ported. I would be surprised if it could already dump native
> > executables for OS X.
>
> SBCL can't dump "native executables" (yet, anyway), but what Paolo
> described was not dumping of a "native executable", but rather
> generating a core (image) file containing the app, bundling the
> runtime executable and running things from a shell script. This can be
> done equally easily with SBCL. Packaging it up as a .app style
> application should also be fairly easy, though I've never written an
> app like that myself so I don't know what might be involved.
There's no rocket science involved, but I wouldn't describe it as easy.
I've done it by assembling the parts of the bundle (the "Foo.app"
directory) by hand. Getting your app to look like an OSX app is a
matter of building the right directory structure and putting the lisp
executable and associated image file in the right places in the
structure, then making a couple of text files that tell OSX where to
look for the executable and other files.
That's the easy part. The more involved part is stuff like arranging to
make the right process handle events, marshaling the right data back
and forth between lisp and theframeworks and so on. Openmcl has
low-level but functional tools for doing all that stuff. SBCL doesn't
yet, though it should be a simple matter of programming given an FFI.
On Mon, 15 Sep 2003 09:43:08 +0100, Christophe Rhodes
<·····@cam.ac.uk> wrote:
>Douglas Philips <····@mac.com> writes:
>
>> What Mac OS X support. Nothing mentioned on any of: http://sbcl.sf.net
>> http://sbcl.sourceforge.net/ports.php (which seems quite old) nd
>> http://sbcl-internals.cliki.net/PPC
>> says:
>> What about MacOS X
>
>Thanks. I've updated these pages to mention Darwin support.
>
>> Ye Gods, why would I even consider SBCL over CLISP? At least I can get
>> CLisp to build, at least I can get an old (2.29) version binaries.
>
>I'd hope that one would consider at least looking at both. Sorry that
>the perception you gained from the various websites you looked at was
>that MacOS X was an unsupported platform.
>
>> Look, I'm not saying that CLisp is perfect, but its rather ironic the
>> invective lobbed at it esp. by comparison with SBCL when SBCL looks to
>> what a newbie may find (see the pages listed above) as dead, and
>> certainly not close for MacOS/X.
>
>I hope that the current situation is better, but if you do find
>another "newbie" route into the sbcl information web which gives the
>impression that the project is dead or that platforms which should be
>mentioned aren't, feedback to the developer lists is always
>appreciated.
>
>Christophe
As an aside, I also had this impression until today. I'm running
CLISP and OpenMCL on my Power book, but I have never installed SBCL,
because the docs did make it appear there was no SBCL port to OS 10.2.
I didn't have any problems getting CLISP installed, it was my very
first attempt at installing any Lisp system and I got it installed and
operating with ILISP via Emacs before I got any other Lisp system
installed.
This is just my perspective as a relative newbie to Lisp.
Doug Tolton
(format t ···@~a~a.~a" "dtolton" "ya" "hoo" "com")
From: Alex McGuire
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <bk6e6e$sqg$1@news-reader3.wanadoo.fr>
Justin Dubs wrote:
>Depends on your platform and financial situation. Given that you are
>a newbie I assume you'll want a free compiler. The main ones seem to
>be:
>
>CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
>OpenMCL http://openmcl.clozure.com/ OS X and Linux PPC
>CMUCL http://www.cons.org/cmucl/ Linux, BSD
>SBCL http://sbcl.sourceforge.net/ Linux, BSD
>
>Hope that helps,
>
>Justin Dubs
>
>
Is there any reason you didn't mention gcl?
From: Christopher C. Stacy
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <upti0hhiy.fsf@dtpq.com>
>>>>> On Tue, 16 Sep 2003 09:17:51 +0100, Alex McGuire ("Alex") writes:
Alex> Justin Dubs wrote:
>> Depends on your platform and financial situation. Given that you are
>> a newbie I assume you'll want a free compiler. The main ones seem to
>> be:
>>
>> CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
>> OpenMCL http://openmcl.clozure.com/ OS X and Linux PPC
>> CMUCL http://www.cons.org/cmucl/ Linux, BSD
>> SBCL http://sbcl.sourceforge.net/ Linux, BSD
>>
>> Hope that helps,
>>
>> Justin Dubs
>>
Alex> Is there any reason you didn't mention gcl?
It's not an ANSI Common Lisp?
>>>>> "Christopher" == Christopher C Stacy <······@dtpq.com> writes:
>>>>> On Tue, 16 Sep 2003 09:17:51 +0100, Alex McGuire ("Alex") writes:
Alex> Justin Dubs wrote:
>>> Depends on your platform and financial situation. Given that you are
>>> a newbie I assume you'll want a free compiler. The main ones seem to
>>> be:
>>>
>>> CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
>>> OpenMCL http://openmcl.clozure.com/ OS X and Linux PPC
>>> CMUCL http://www.cons.org/cmucl/ Linux, BSD
>>> SBCL http://sbcl.sourceforge.net/ Linux, BSD
>>>
>>> Hope that helps,
>>>
>>> Justin Dubs
>>>
Alex> Is there any reason you didn't mention gcl?
Christopher> It's not an ANSI Common Lisp?
Neither is clisp. But both clisp and gcl are trying to be more ANSI.
Ray
From: Christopher C. Stacy
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <ufzive4p2.fsf@dtpq.com>
>>>>> On Tue, 16 Sep 2003 16:26:08 -0400, Raymond Toy ("Raymond") writes:
>>>>> "Christopher" == Christopher C Stacy <······@dtpq.com> writes:
>>>>> On Tue, 16 Sep 2003 09:17:51 +0100, Alex McGuire ("Alex") writes:
Alex> Justin Dubs wrote:
>>>> Depends on your platform and financial situation. Given that you are
>>>> a newbie I assume you'll want a free compiler. The main ones seem to
>>>> be:
>>>>
>>>> CLISP http://clisp.sourceforge.net/ Linux, BSD, Win32, OS X
>>>> OpenMCL http://openmcl.clozure.com/ OS X and Linux PPC
>>>> CMUCL http://www.cons.org/cmucl/ Linux, BSD
>>>> SBCL http://sbcl.sourceforge.net/ Linux, BSD
>>>>
>>>> Hope that helps,
>>>>
>>>> Justin Dubs
>>>>
Alex> Is there any reason you didn't mention gcl?
Christopher> It's not an ANSI Common Lisp?
Raymond> Neither is clisp. But both clisp and gcl are trying to be more ANSI.
Last time I looked, CLISP was credibly close to ANSI Common Lisp,
and was commonly used by people as an ANSI Common Lisp;
GCL was nowhere near that situation.
Has that changed?
>>>>> "Christopher" == Christopher C Stacy <······@dtpq.com> writes:
>>>>> On Tue, 16 Sep 2003 16:26:08 -0400, Raymond Toy ("Raymond") writes:
>>>>> "Christopher" == Christopher C Stacy <······@dtpq.com> writes:
Christopher> It's not an ANSI Common Lisp?
Raymond> Neither is clisp. But both clisp and gcl are trying to be more ANSI.
Christopher> Last time I looked, CLISP was credibly close to ANSI Common Lisp,
Christopher> and was commonly used by people as an ANSI Common Lisp;
Christopher> GCL was nowhere near that situation.
Christopher> Has that changed?
I think the GCL folks are working hard to make it ANSI. At least
there's a compile time option for ANSI. But I don't really use gcl,
so I can't say how close it is.
Ray
From: Paul F. Dietz
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <49ucnX6-cPygbvWiXTWJhA@dls.net>
Raymond Toy wrote:
> I think the GCL folks are working hard to make it ANSI.
There's a goal to make it ANSI, but the project could use
some more people to hack at the growing list of known
compliance issues.
Paul
"craig dicker" <·······@··············@hn.ozemail.com.au> wrote in message news:<·····················@nnrp1.ozemail.com.au>...
lisp is generally interpreted. librep is a good hybrid dialect of
lisp, which has many cool features like modules (ala certain scheme
dialects) and asynchronous timers, and is embeddable through C. GNU
Emacs lisp is also very good. clisp does common lisp, which is useful
for system programming. umb-scheme and guile might be worth trying for
extensions, but seem to lack certain desirable features for bigger
tasks.
From: Peter Seibel
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <m3he38o2g0.fsf@javamonkey.com>
·············@yahoo.com (Mr. Berserker) writes:
> "craig dicker" <·······@··············@hn.ozemail.com.au> wrote in message news:<·····················@nnrp1.ozemail.com.au>...
>
> lisp is generally interpreted.
That's not correct (and hasn't been for probably several decades now.)
-Peter
--
Peter Seibel ·····@javamonkey.com
Lisp is the red pill. -- John Fraser, comp.lang.lisp
From: Edi Weitz
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <878yokxwu1.fsf@bird.agharta.de>
Just in case someone might read this and think it is true:
On 19 Sep 2003 13:52:40 -0700, ·············@yahoo.com (Mr. Berserker) wrote:
> lisp is generally interpreted.
No, it isn't. Common Lisp, the main Lisp dialect in use nowadays, is
generally compiled to machine code. The only exception I know of is
CLISP which compiles to byte code similar to Java.
> clisp does common lisp, which is useful for system programming.
Common Lisp is a general purpose programming language which is useful
for almost anything you can think of. In fact, there are some people
(not me) who think that it is useful for anything _expect_ systems
programming.
Edi.
From: Edi Weitz
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <874qz8xwqr.fsf@bird.agharta.de>
On 20 Sep 2003 00:00:38 +0200, Edi Weitz <···@agharta.de> wrote:
> > clisp does common lisp, which is useful for system programming.
>
> Common Lisp is a general purpose programming language which is
> useful for almost anything you can think of. In fact, there are some
> people (not me) who think that it is useful for anything _expect_
> systems programming.
^^^^^^^^
That should have been "except", of course... :)
In article <··············@bird.agharta.de>, Edi Weitz <···@agharta.de>
wrote:
> On 20 Sep 2003 00:00:38 +0200, Edi Weitz <···@agharta.de> wrote:
>
> > > clisp does common lisp, which is useful for system programming.
> >
> > Common Lisp is a general purpose programming language which is
> > useful for almost anything you can think of. In fact, there are some
> > people (not me) who think that it is useful for anything _expect_
> > systems programming.
>
> ^^^^^^^^
>
> That should have been "except", of course... :)
Try a Lisp Machine sometime.
From: Edi Weitz
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <87y8wkwgen.fsf@bird.agharta.de>
On Sat, 20 Sep 2003 00:35:41 +0200, Rainer Joswig <······@lispmachine.de> wrote:
> Try a Lisp Machine sometime.
Wanna lend me one of yours? I could jump in my car and get it tomorrow
morning when I buy rolls. I just got a copy of "Lisp Lore" so it'd be
a perfect fit... :)
Cheers,
Edi.
Mr. Berserker:
> craig dicker:
> lisp is generally interpreted.
The Lisp project started in 1958, and an explicit goal was to build a
compiler for the new language. It was a lucky break that they had an
interpreter for s-expression Lisp before they had a compiler. In 1959
or 1960 they had a compiler for the very first Lisp - Lisp1 - written
in Lisp.
--
"Hurry if you still want to see something. Everything is vanishing."
-- Paul C�zanne (1839-1906)
·············@yahoo.com (Mr. Berserker) wrote in message news:<····························@posting.google.com>...
> "craig dicker" <·······@··············@hn.ozemail.com.au> wrote in message news:<·····················@nnrp1.ozemail.com.au>...
>
> lisp is generally interpreted. librep is a good hybrid dialect of
> lisp, which has many cool features like modules (ala certain scheme
> dialects) and asynchronous timers, and is embeddable through C. GNU
> Emacs lisp is also very good. clisp does common lisp, which is useful
> for system programming. umb-scheme and guile might be worth trying for
> extensions, but seem to lack certain desirable features for bigger
> tasks.
Just to further correct the misleading information in this post,
umb-scheme seems to one of the deadest schemes out there. Guile is
also not the most popular choice among scheme users.
--
G.
········@pithekos.net (Grzegorz Chrupala) writes:
> ·············@yahoo.com (Mr. Berserker) wrote in message news:<····························@posting.google.com>...
> > "craig dicker" <·······@··············@hn.ozemail.com.au> wrote in message news:<·····················@nnrp1.ozemail.com.au>...
> >
> > lisp is generally interpreted. librep is a good hybrid dialect of
> > lisp, which has many cool features like modules (ala certain scheme
> > dialects) and asynchronous timers, and is embeddable through C. GNU
> > Emacs lisp is also very good. clisp does common lisp, which is useful
> > for system programming. umb-scheme and guile might be worth trying for
> > extensions, but seem to lack certain desirable features for bigger
> > tasks.
>
> Just to further correct the misleading information in this post,
> umb-scheme seems to one of the deadest schemes out there. Guile is
> also not the most popular choice among scheme users.
And to correct another misleading information, of the Common-Lisp
implementations I know, the majority is compiled, not interpreted:
- sbcl -- pure compiled
- cmucl -- pure compiled
- gcl -- pure compiled
- ecl -- bytecode compiler and interpreter
- clisp -- bytecode compiler and interpreter
--
__Pascal_Bourguignon__
http://www.informatimago.com/
Do not adjust your mind, there is a fault in reality.
From: Paul F. Dietz
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <b_OdnXRFY92UKvGiXTWJiw@dls.net>
Pascal Bourguignon wrote:
> - cmucl -- pure compiled
> - gcl -- pure compiled
These are erroneous. In particular, in gcl, compilation is a rather
heavyweight procedure (going through gcc) so you wouldn't want to do it for
a simple eval.
Paul
In article <··············@thalassa.informatimago.com>,
Pascal Bourguignon <····@thalassa.informatimago.com> wrote:
> ········@pithekos.net (Grzegorz Chrupala) writes:
>
> > ·············@yahoo.com (Mr. Berserker) wrote in message news:<····························@posting.google.com>...
> > > "craig dicker" <·······@··············@hn.ozemail.com.au> wrote in message news:<·····················@nnrp1.ozemail.com.au>...
> > >
> > > lisp is generally interpreted. librep is a good hybrid dialect of
> > > lisp, which has many cool features like modules (ala certain scheme
> > > dialects) and asynchronous timers, and is embeddable through C. GNU
> > > Emacs lisp is also very good. clisp does common lisp, which is useful
> > > for system programming. umb-scheme and guile might be worth trying for
> > > extensions, but seem to lack certain desirable features for bigger
> > > tasks.
> >
> > Just to further correct the misleading information in this post,
> > umb-scheme seems to one of the deadest schemes out there. Guile is
> > also not the most popular choice among scheme users.
>
> And to correct another misleading information, of the Common-Lisp
> implementations I know, the majority is compiled, not interpreted:
>
> - sbcl -- pure compiled
> - cmucl -- pure compiled
> - gcl -- pure compiled
> - ecl -- bytecode compiler and interpreter
> - clisp -- bytecode compiler and interpreter
Have you checked your facts???
> lisp
CMU Common Lisp CVS release-18e-branch + minimal debian patches, running on do
With core: /usr/lib/cmucl/lisp.core
Dumped on: Sat, 2003-09-20 09:18:33+02:00
For support see http://www.cons.org/cmucl/support.html Send bug reports to the debian BTS.
or to ········@debian.org
type (help) for help, (quit) to exit, and (demo) to see the demos
Loaded subsystems:
Python 1.1, target Intel x86
CLOS 18e (based on PCL September 16 92 PCL (f))
* (defun foo () 'bar)
FOO
* (describe #'foo)
#<Interpreted Function FOO {4801E949}> is function.
Arguments:
There are no arguments.
Its defined argument types are:
NIL
Its result type is:
*
Its definition is:
(LAMBDA () (BLOCK FOO 'BAR))
*
So CMUCL provides an interpreter, too. Atleast at some point
in time CMUCL also had a byte-code compiler.
I guess GCL has an interpreter, too. Let's see:
> gcl
GCL (GNU Common Lisp) (2.6.1) Tue Sep 16 15:08:40 UTC 2003
Licensed under GNU Library General Public License
Dedicated to the memory of W. Schelter
Use (help) to get some basic information on how to use GCL.
>(defun foo () 'bar)
FOO
>(describe #'foo)
(LAMBDA-BLOCK FOO () 'BAR) - function
>(compile 'foo)
Compiling gazonk0.lsp.
End of Pass 1.
End of Pass 2.
OPTIMIZE levels: Safety=0 (No runtime error checking), Space=0, Speed=3
Finished compiling gazonk0.lsp.
Loading gazonk0.o
start address -T 0x864af60 Finished loading gazonk0.o
#<compiled-function FOO>
>(describe #'foo)
#<compiled-function FOO> - COMPILED-FUNCTION
>
okay. Interpreter + compiler, too.
CMUCL 18d is not "pure" compiled.
For example:
CMU Common Lisp 18d, running on cawtech.freeserve.co.uk
Send questions and bug reports to your local CMU CL maintainer, or to ··········@cons.org. and ·········@cons.org. respectively.
Loaded subsystems:
Python 1.0, target Intel x86
CLOS based on PCL version: September 16 92 PCL (f)
* (defun fib (n)
(cond ((zerop n) 1)
((= 1 n) 1)
(t (+ (fib (- n 1))
(fib (- n 2))))))
FIB
* (compiled-function-p #'fib)
T
* (time (fib 23))
Compiling LAMBDA NIL:
Compiling Top-Level Form:
Evaluation took:
7.8 seconds of real time
0.0 seconds of user run time
7.784829 seconds of system run time
0 page faults and
11443944 bytes consed.
46368
* (describe 'fib)
FIB is an internal symbol in the COMMON-LISP-USER package.
Function: #<Interpreted Function FIB {48017799}>
Function arguments:
(N)
Its defined argument types are:
(T)
Its result type is:
*
Its definition is:
(LAMBDA (N) (BLOCK FIB #))
* (compile 'fib)
Compiling LAMBDA (N):
Compiling Top-Level Form:
FIB
NIL
NIL
* (describe 'fib)
FIB is an internal symbol in the COMMON-LISP-USER package.
Function: #<Function FIB {48B52611}>
Function arguments:
(n)
Its defined argument types are:
(T)
Its result type is:
*
On Saturday, 9/20/03 10:07:18 pm [+1] it was compiled from:
#(#'(LAMBDA # #))
* (time (fib 23))
Compiling LAMBDA NIL:
Compiling Top-Level Form:
Evaluation took:
0.06 seconds of real time
0.0 seconds of user run time
0.054613 seconds of system run time
0 page faults and
0 bytes consed.
46368
I don't know what is happening with
(compiled-function-p #'fib) => t
It is definitely interpreted.
I compile it and it runs *much* faster.
The reason I insist on this picky little point is that if
you are told that CMUCL is *pure* compiled, and you manage
to bypass the compiler by typing a function definition
straight to the REPL instead of loading the file containing,
you will not realise it was interpreted, and you will come
away from the experience believing the myth that Lisp is
slow.
Is there a good reason why CMUCL exposes its interpreter? I
thought it might be doing it so that if you have an inner
loop consing up lots of one off functions for eval, you can
avoid the cost of compiling throw away code, but trying it
out, eval compiles its function definitions.
What happens with 18e?
Alan Crowe
Edinburgh
Scotland
From: Christophe Rhodes
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <sqvfrmmmie.fsf@lambda.jcn.srcf.net>
····@cawtech.freeserve.co.uk (Alan S. Crowe) writes:
> CMUCL 18d is not "pure" compiled.
> [snip]
> What happens with 18e?
Exactly the same.
COMPILED-FUNCTION-P does not ask "has this function been compiled to
machine code?"; it asks "has this function been minimally-compiled?".
Loosely, the second question is essentially "have all the macros been
expanded?".
Internally, CMUCL's interpreter (again, I'm going to speak loosely)
walks the highest-level internal representation of the code, and not
the s-expr tree itself. So it is interpreting a compiled (in the
Standard sense, not the standard sense :-) representation.
Christophe
--
http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)
From: Eric Marsden
Subject: Re: hi all i am new to this so what is a good compiler to use for lisp
Date:
Message-ID: <wzibrtep0ul.fsf@melbourne.laas.fr>
>>>>> "pb" == Pascal Bourguignon <····@thalassa.informatimago.com> writes:
pb> - cmucl -- pure compiled
For the record, CMUCL includes a native-code compiler, a byte-code
compiler, and an interpreter (in fact two interpreters -- see the
document above for the details). Which evaluation mode is used depends
on how you interact with CMUCL.
It's disappointing to see people post incorrect information here,
particularly considering how easy it is to find out the truth.
The top hit on Google for "CMUCL compiler interpreter" is
<URL:http://www.cons.org/cmucl/doc/different-compilers.html>
--
Mankind have a great aversion to intellectual labor; but even
supposing knowledge to be easily attainable, more people would be
content to be ignorant than would take even a little trouble to
acquire it. -- Samuel "STFW" Johnson