Hello,
This post is a mix of questions and suggestions; as it deals specifically
with CMUCL, I hope it is not too specific for this list, though having seen
the names of some of the maintainers as posters, I hope they should be able
to use this.
I'm new to Lisp. In trying to select a free compiler, I've downloaded
clisp, Steel-Bank and CMU CL. Not having any prior Lisp installation, I
went for the binary in each case (my Linux distribution comes with about 4
different Prolog implementations, but no Lisp...). I was able to install
both clisp and Steel-Bank very easily, but CMUCL proves a bit more
tricky.
First thing, selecting which CMUCL package to download off
www.cons.org/cmucl/ is a little disconcerting. Under "Getting current CMU
CL binaries", which is said to contain "Binaries built from source _more
recent than the latest release_", one finds the 18a version, from 1997,
whereas an 18c version is to be found under "Getting the [supposedly
obsolete] latest CMU CL release". I'd suggest clarifying things.
The cmucl-18c.x86.linux.glibc22.tgz package I opened (tar xvfz) under
/usr/local/src/ to make sure I'd see the whole content in one place. Sure
enough, it opens in a set of files and directories without a common root
(which I consider the "polite" thing to do). I reopened if from /usr/local
and the files went various ways (bin, lib, man), but apparently not quite
the way they should, however that is. When I run 'lisp', it tells me it
cannot open file "/usr/lib/cmucl/lisp.core". So the tar file should really
be opened under /usr/ rather than /usr/local/ ? Well, that's not what the
man page says; it says the default is /usr/local/lib/cmucl/lib. But then
again, that's not where the file went; lisp.core is to be found under
/usr/local/lib directly.
Puzzling. There is copious documentation on the internals of CMU CL, but I
couldn't find a README or INSTALL that would help.
So the question is, what is the proper way to open the tar file ?
And the suggestion for the maintainers would be to please copy the
Steel-Bank (hope I'm not in war zone here) installation process: README,
INSTALL, and install.sh script from something like a cmucl-18c root
directory.
I've read fine things on the compiler and am looking forward to using it.
Thanks,
Fran�ois B�dard
>>>>> "Fran�ois" == Fran�ois B�dard <bedard @ info.uqam.ca> writes:
Fran�ois> The cmucl-18c.x86.linux.glibc22.tgz package I opened (tar xvfz) under
Fran�ois> /usr/local/src/ to make sure I'd see the whole content in one place. Sure
Fran�ois> enough, it opens in a set of files and directories without a common root
Fran�ois> (which I consider the "polite" thing to do). I reopened if from /usr/local
Fran�ois> and the files went various ways (bin, lib, man), but apparently not quite
Fran�ois> the way they should, however that is. When I run 'lisp', it tells me it
Fran�ois> cannot open file "/usr/lib/cmucl/lisp.core". So the tar file should really
Fran�ois> be opened under /usr/ rather than /usr/local/ ? Well, that's not what the
Fran�ois> man page says; it says the default is /usr/local/lib/cmucl/lib. But then
Fran�ois> again, that's not where the file went; lisp.core is to be found under
Fran�ois> /usr/local/lib directly.
The linux version has /usr/lib/cmucl/lisp.core as the default. The
manpage is way out-of-date.
Look at the sample-wrapper that came with it. Place it wherever you
want it, and probably rename it to cmucl or cmulisp or lisp or
whatever you want. Edit the contents appropriately, and you'll be
set.
Fran�ois> Puzzling. There is copious documentation on the internals of CMU CL, but I
Fran�ois> couldn't find a README or INSTALL that would help.
Fran�ois> So the question is, what is the proper way to open the tar file ?
You did the right thing. Just edit sample-wrapper appropriately and
away you go.
Fran�ois> And the suggestion for the maintainers would be to please copy the
Fran�ois> Steel-Bank (hope I'm not in war zone here) installation process: README,
Fran�ois> INSTALL, and install.sh script from something like a cmucl-18c root
Fran�ois> directory.
A deficiency. Hopefully it will be fixed before the next release
whenver that may be.
Fran�ois> I've read fine things on the compiler and am looking forward to using it.
Since you already have SBCL working, you probably won't care about the
minor differences between SBCL and CMUCL.
Ray
Raymond Toy wrote:
>>>>>> "Fran�ois" == Fran�ois B�dard <bedard @ info.uqam.ca> writes:
>
> Fran�ois> [describes his experience trying to settle where to untar
> Fran�ois> the distribution.]
>
> The linux version has /usr/lib/cmucl/lisp.core as the default. The
> manpage is way out-of-date.
>
> Look at the sample-wrapper that came with it. Place it wherever you
> want it, and probably rename it to cmucl or cmulisp or lisp or
> whatever you want. Edit the contents appropriately, and you'll be
> set.
>
> Fran�ois> I couldn't find a README or INSTALL that would help.
>
> Fran�ois> So the question is, what is the proper way to open the tar
> file ?
>
> You did the right thing. Just edit sample-wrapper appropriately and
> away you go.
Thanks. Guess I needed some hand-holding there.
>
> Fran�ois> And the suggestion for the maintainers would be to please
> Fran�ois> copy the Steel-Bank (hope I'm not in war zone here)
> Fran�ois> installation process: README, INSTALL, and install.sh script
> Fran�ois> from something like a cmucl-18c root directory.
>
> A deficiency. Hopefully it will be fixed before the next release
> whenver that may be.
"Install" is apparently more clueless-proof than "sample-wrapper"...
>
> Fran�ois> I've read fine things on the compiler and am looking forward
> to using it.
>
> Since you already have SBCL working, you probably won't care about the
> minor differences between SBCL and CMUCL.
>
Somewhat unfortunate that there are two if they're so close. I understand
CMUCL is the long time original whereas SBCL is more portable; are there
plans on merging ?
> Ray
>
Hello!
In article <··············@rtp.ericsson.se>,
Raymond Toy <···@rtp.ericsson.se> wrote:
>[...]
>Since you already have SBCL working, you probably won't care about the
>minor differences between SBCL and CMUCL.
For the Lisp core, that will probably be true. However if there's an
interest in CMUCL's extensions, you'll mostly find them missing
for sbcl.
>Ray
Kind regards,
Hannah.
>>>>> "Hannah" == Hannah Schroeter <······@schlund.de> writes:
Hannah> Hello!
Hannah> In article <··············@rtp.ericsson.se>,
Hannah> Raymond Toy <···@rtp.ericsson.se> wrote:
>> [...]
>> Since you already have SBCL working, you probably won't care about the
>> minor differences between SBCL and CMUCL.
Hannah> For the Lisp core, that will probably be true. However if there's an
Hannah> interest in CMUCL's extensions, you'll mostly find them missing
Hannah> for sbcl.
But he said he was new to lisp. What would he want with extensions?
I'm not new to lisp and I don't use any extensions, not even sockets.
(I do depend on a smart compiler, however.)
Ray
Hello!
In article <··············@rtp.ericsson.se>,
Raymond Toy <···@rtp.ericsson.se> wrote:
>[...]
>But he said he was new to lisp. What would he want with extensions?
>I'm not new to lisp and I don't use any extensions, not even sockets.
>(I do depend on a smart compiler, however.)
I'm definitely not a CL expert, however, I would be interested
in playing with Hemlock, and I would definitely be interested in
e.g. Sockets, and if necessary, threads. A good part of my work
at my company is writing some network services, and I'd really like
to try writing them in Lisp instead of C/C++. (However, I've written
some few in Haskell/GHC and Ocaml, already, and that was quite a bit
nicer than C/C++, too.)
Kind regards,
Hannah.
······@schlund.de (Hannah Schroeter) writes:
> In article <··············@rtp.ericsson.se>,
> Raymond Toy <···@rtp.ericsson.se> wrote:
> >[...]
>
> >Since you already have SBCL working, you probably won't care about the
> >minor differences between SBCL and CMUCL.
>
> For the Lisp core, that will probably be true. However if there's an
> interest in CMUCL's extensions, you'll mostly find them missing
> for sbcl.
- All the socket stuff (internet.lisp, WIRE, REMOTE) is gone, as are
big chunks of the Unix interface.
- No Hemlock, afaik no CLIO, CLUE
+ PCL counts as "core" not "extension"; that's all there
+ The Alien interface (general FFI) is still there
* internet.lisp has been ported and is available "unbundled"
* as is the other socket interface db-sockets (though it may be out of
date as I haven't looked at it recently)
* A CLX port is also available separately
* mk-defsystem can be had from CLOCC or by installing the
Common-Lisp-Controllerised binaries in CCLAN (see below, see below)
* Recent ILISP supports SBCL
You can probably find pointers to most of the unbundled bits by
looking at
http://sbcl.sourceforge.net/libs.php
http://ww.telent.net/cliki/SBCL
http://www.geocrawler.com/lists/3/SourceForge/1663/0/ (mailing list archives)
http://clocc.sourceforge.net/
Work is underway to collect these kinds of thing together (for CMUCL
as well as for SBCL, and eventually we hope for other implementations
too) in some kind of organised fashion, but it's early days yet. See
http://ww.telent.net/cliki/cclan
Incidentally, the next release of SBCL should include the Alpha
compiler backend (transplanted from CMUCL) and runtime support for
Linux/Alpha. SBCL will no longer be x86-only ...
-dan
--
http://ww.telent.net/cliki/ - Link farm for free CL-on-Unix resources
[CC ·········@lists.sourceforge.net because of the SBCL specific problem]
In article <··············@noetbook.telent.net>,
Daniel Barlow <···@telent.net> wrote:
>[...]
>Incidentally, the next release of SBCL should include the Alpha
>compiler backend (transplanted from CMUCL) and runtime support for
>Linux/Alpha. SBCL will no longer be x86-only ...
Yes, but as "compensation" it seems that it doesn't work on BSD,
especially OpenBSD, anymore.
I have 0.6.11.41 running, and newer sources from the CVS fail to compile
with that host, especially in the C runtime system.
There are problems that the source uses the names "st_mtime" ...
in an own structure to map the OS specific struct stat to a common
structure, however, on OpenBSD, struct stat contains:
...
struct timespec st_atimespec; /* time of last access */
struct timespec st_mtimespec; /* time of last data modification */
struct timespec st_ctimespec; /* time of last file status change */
...
and defines:
#define st_atime st_atimespec.tv_sec
[...]
#define st_mtime st_mtimespec.tv_sec
[...]
#define st_ctime st_ctimespec.tv_sec
src/runtime/wrap.c declares structure members
time_t st_atime; /* time of last access */
time_t st_mtime; /* time of last modification */
time_t st_ctime; /* time of last change */
which get expanded to
time_t st_atimespec.tv_sec;
time_t st_mtimespec.tv_sec;
time_t st_ctimespec.tv_sec;
which is a syntax error.
Once, I tried to fix it and did compile the runtime however got a strange
error a bit later when compiling some Lisp file.
Perhaps some SBCL person might look into that?
Kind regards,
Hannah.
>>>>> " fb" == Fran�ois B�dard <bedard @ info.uqam.ca> writes:
fb> Puzzling. There is copious documentation on the internals of
fb> CMU CL, but I couldn't find a README or INSTALL that would
fb> help.
it's true that CMUCL is a little lacking in this respect. The easiest
way to install on Linux is to use the Debian packages (converting to
RPMs using alien if necessary).
There are some more detailed guidelines for installing from the
tarballs at <URL:http://www.laas.fr/~emarsden/cons.org/download.html>
(these web pages should one day end up at cons.org). Let me know if
these instructions can be improved.
--
Eric Marsden <URL:http://www.laas.fr/~emarsden/>
Eric Marsden wrote:
>>>>>> " fb" == Fran�ois B�dard <bedard @ info.uqam.ca> writes:
>
> fb> Puzzling. There is copious documentation on the internals of
> fb> CMU CL, but I couldn't find a README or INSTALL that would
> fb> help.
>
> it's true that CMUCL is a little lacking in this respect. The easiest
> way to install on Linux is to use the Debian packages (converting to
> RPMs using alien if necessary).
>
> There are some more detailed guidelines for installing from the
> tarballs at <URL:http://www.laas.fr/~emarsden/cons.org/download.html>
> (these web pages should one day end up at cons.org). Let me know if
> these instructions can be improved.
Btw. Why do you wait submitting them to cons.org?
Regards,
Jochen
Eric Marsden wrote:
>>>>>> " fb" == Fran�ois B�dard <bedard @ info.uqam.ca> writes:
>
> fb> There is copious documentation on the internals of
> fb> CMU CL, but I couldn't find a README or INSTALL that would
> fb> help.
>
> it's true that CMUCL is a little lacking in this respect. The easiest
> way to install on Linux is to use the Debian packages (converting to
> RPMs using alien if necessary).
>
> There are some more detailed guidelines for installing from the
> tarballs at <URL:http://www.laas.fr/~emarsden/cons.org/download.html>
> (these web pages should one day end up at cons.org). Let me know if
> these instructions can be improved.
>
I see the page has been modified as of yesterday. Don't know what it looked
like before but it is now a model in clarity. I'd merely suggest either
including it with the release or linking to it from cons.org. Or, better
still, as Jochen Schmidt suggests in a follow-up to yours, replacing the
pages at cons.org with those on your site.
In article <·······················@carnaval.risq.qc.ca>, Fran�ois B�dard wrote:
> Hello,
>
> This post is a mix of questions and suggestions; as it deals specifically
> with CMUCL, I hope it is not too specific for this list, though having seen
> the names of some of the maintainers as posters, I hope they should be able
> to use this.
>
> I'm new to Lisp. In trying to select a free compiler, I've downloaded
> clisp, Steel-Bank and CMU CL. Not having any prior Lisp installation, I
> went for the binary in each case (my Linux distribution comes with about 4
> different Prolog implementations, but no Lisp...). I was able to install
> both clisp and Steel-Bank very easily, but CMUCL proves a bit more
> tricky.
If you're installing on a redhat system, I've got some very simple rpms that
will at least get you started. You can get them at http://www.caddr.com/cmucl.
Please note that I made these myself and they're not endorsed by or supported
by any of the cmucl volunteers or by anyone at cons.org.
--
miles egan
Miles Egan wrote:
> Fran�ois B�dard wrote:
>> I was able to install both clisp and Steel-Bank very easily,
>> but CMUCL proves a bit more tricky.
>
> If you're installing on a redhat system, I've got some very simple rpms
> that will at least get you started. You can get them at
> http://www.caddr.com/cmucl. Please note that I made these myself and
> they're not endorsed by or supported by any of the cmucl volunteers or by
> anyone at cons.org.
>
Thanks. I'm actually running Mandrake, which is supposedly "99%" compatible
with Red Hat, but then again... Actually, the binary distribution has just
a few files and I can always shuffle them around, but since I prefer to
install my non-distribution applications under /usr/local than under /usr/
I was wondering whether there was a work-around for placing lisp.core under
/usr/lib/cmucl/ as the program will have it. It was suggested to me
privately I should use symlinks - should have thought of that. Also thought
I'd file some sort of a low-priority "bug" report.
>>>>> "FB" == Fran�ois B�dard <bedard @ info.uqam.ca> writes:
FB> [...] I was wondering whether there was a work-around for placing
FB> lisp.core under /usr/lib/cmucl/ as the program will have it.
Try setting the environment variable CMUCLLIB, as is demonstrated in the
sample-wrapper script, if that is shipped with the linux binaries.
TTFN,
'mr
--
Martin Rydstr|m, ·····@cd.chalmers.se ; rationalize till I'm blue in the face,
; you cannot lose if you throw the race.
Looking for a LispM to play with. Any and all offers/ideas appreciated.
Also - " - job - " - - " -
On Fri, 04 May 2001 15:56:57 GMT, "Fran�ois B�dard" <bedard @ info.uqam.ca>
wrote:
> different Prolog implementations, but no Lisp...). I was able to install
> both clisp and Steel-Bank very easily, but CMUCL proves a bit more
> tricky.
The EncyCMUCLopedia (see my signature below) also includes some
instructions and introductory documentation on installing CMU CL.
> First thing, selecting which CMUCL package to download off
> www.cons.org/cmucl/ is a little disconcerting. Under "Getting current CMU
This site is more up to date:
http://www.laas.fr/~emarsden/cons.org/
Paolo
--
EncyCMUCLopedia * Extensive collection of CMU Common Lisp documentation
http://cvs2.cons.org:8000/cmucl/doc/EncyCMUCLopedia/
Paolo Amoroso <·······@mclink.it> writes:
> On Fri, 04 May 2001 15:56:57 GMT, "Fran�ois B�dard" <bedard @ info.uqam.ca>
> wrote:
>
> > different Prolog implementations, but no Lisp...). I was able to install
> > both clisp and Steel-Bank very easily, but CMUCL proves a bit more
> > tricky.
>
> The EncyCMUCLopedia (see my signature below) also includes some
> instructions and introductory documentation on installing CMU CL.
That's what worked the best for me. read
doc/devenv/installing-18c-on-rh70.txt
--
Brady Montz
······@balestra.org
Paolo Amoroso wrote:
> On Fri, 04 May 2001 15:56:57 GMT, "Fran�ois B�dard" <bedard @
> info.uqam.ca> wrote:
>
>> different Prolog implementations, but no Lisp...). I was able to install
>> both clisp and Steel-Bank very easily, but CMUCL proves a bit more
>> tricky.
>
> The EncyCMUCLopedia (see my signature below) also includes some
> instructions and introductory documentation on installing CMU CL.
>
I've downloaded all 3+ (zipped) megs of it. That should provide reading for
years to come :), though I might first make it through Graham's.
>
>> First thing, selecting which CMUCL package to download off
>> www.cons.org/cmucl/ is a little disconcerting. Under "Getting current CMU
>
> This site is more up to date:
>
> http://www.laas.fr/~emarsden/cons.org/
>
It is indeed more informative and has even - as of 3 May - a link to rpms
at rpmfind.net. May I note though that ALU still links to cons.org.
>
> Paolo
Fran�ois