I've installed ILisp, copied ilisp.emacs to .emacs and tweaked the
various paths within. But when I ruyn emacs and do "M-x cmulisp",
visit a foo.lisp file and then try to evaluate (/ 1 0) I get the elisp
stack trace, not the one from cmulisp.
How do I force ILisp and CMUCL to start working? All the docs I can
find on the net seem a little light on this point.
On 5 Nov 2003 12:28:35 -0800, ······@mail.com (KanZen) wrote:
> I've installed ILisp, copied ilisp.emacs to .emacs and tweaked the
> various paths within. But when I ruyn emacs and do "M-x cmulisp",
> visit a foo.lisp file and then try to evaluate (/ 1 0) I get the
> elisp stack trace, not the one from cmulisp.
>
> How do I force ILisp and CMUCL to start working? All the docs I can
> find on the net seem a little light on this point.
Have you tried these (both by Bill Clementson)?
<http://cl-cookbook.sourceforge.net/windows.html>
<http://cl-cookbook.sourceforge.net/emacs-ide.html>
Edi.
From: Lupo LeBoucher
Subject: Re: How does ILisp/HyperSpec work, and a general observation about Lisp culture
Date:
Message-ID: <OKydnUeMUe8hLjSiRVn-ug@io.com>
In article <··············@bird.agharta.de>, Edi Weitz <···@agharta.de> wrote:
>On 5 Nov 2003 12:28:35 -0800, ······@mail.com (KanZen) wrote:
>
>> I've installed ILisp, copied ilisp.emacs to .emacs and tweaked the
>> various paths within. But when I ruyn emacs and do "M-x cmulisp",
>> visit a foo.lisp file and then try to evaluate (/ 1 0) I get the
>> elisp stack trace, not the one from cmulisp.
>>
>> How do I force ILisp and CMUCL to start working? All the docs I can
>> find on the net seem a little light on this point.
>
>Have you tried these (both by Bill Clementson)?
>
> <http://cl-cookbook.sourceforge.net/windows.html>
> <http://cl-cookbook.sourceforge.net/emacs-ide.html>
I've tried both. They work for firing up Ilisp. In fact, so does the
code frags in the Ilisp INSTALLATION file.
Here is what sucks: the HyperSpec thingee doesn't do anything. It claims
'cons' and '+' isn't a part of the HyperSpec, which I know to be untrue.
Under other compilation conditions, it sends cryptic elisp debugger
messages to the minibuffer thingee.
I have tried the one which comes with SLIME as well (BTW, SLIME is way
neat -it is not as featureful as Ilisp, and it don't work with clisp-hs
but I like it better). No beer. I have tried different versions of xemacs
and emacs; no beer. One of our group claims to have it running on his
CygWin installation, but it doesn't work on Linux.
I'm actually running a study group: we're going through Paul Graham's
'ANSI Common Lisp' book & so far, the chosen environment is
clisp/ilisp/xemacs (with the option to look at Franz, cmu-cl & slime).
We're a fairly elite group of 13 professional coders and Ph.D.'s in the
hard sciences. Not a dummy in the group; we all intend to use Common Lisp
for some serious business that is difficult or impossible to do in C.
But, we're all having difficulties with the "friendliness" of the
available tools and IDEs. My standard trick of searching the deja archives
for problems does not work either, as nobody every seems to *learn* lisp
using these tools -they just seem to come with an inborn ability to write
code in it or something.
Yes, I have tried Hemlock; it was amusing in that, the documentation for
it doesn't tell you anything useful, like how to fire it up (yes, I
eventually figured out how to) or get it to do anything you might like to
with a text editor, such as open a file. Totally worthless, unless you've
been using it for the last 20 years.
I suspect this language would get more attention if, say, there were a
halfway decent development environment for it, or "how to get started"
type files, such as exist for every other damned thing. For all the
advertised "wow, you can prototype code really fast in this way," aspects
of Lisp I keep hearing about (all of which apply to Python, in my
experience), it sure is a pain in the ass to get the minimal IDE
documentation crapola working.
-Lupo <··@pentagon.io.com>
"Java: the elegant simplicity of C++ and the blazing speed of Smalltalk"
From: Alan Shutko
Subject: Re: How does ILisp/HyperSpec work, and a general observation about Lisp culture
Date:
Message-ID: <87ad7ajgz7.fsf@wesley.springies.com>
··@io.com (Lupo LeBoucher) writes:
> experience), it sure is a pain in the ass to get the minimal IDE
> documentation crapola working.
You would find more answers to your questions if you wrote better
questions. FWIW, I have ilisp working both under Linux with CMUCL
(Debian did it all for me) and under Windows with a Windows CLISP.
If you explain more precisely what you're trying, and what exactly it
does, someone will be able to help you. Check out
http://www.catb.org/~esr/faqs/smart-questions.html if you're having
trouble framing your question.
--
Alan Shutko <···@acm.org> - I am the rocks.
Looking for your cat? Talk to my 300 pound canary.
From: Thomas F. Burdick
Subject: Re: How does ILisp/HyperSpec work, and a general observation about Lisp culture
Date:
Message-ID: <xcvvfpyyned.fsf@famine.OCF.Berkeley.EDU>
··@io.com (Lupo LeBoucher) writes:
> Yes, I have tried Hemlock; it was amusing in that, the documentation for
> it doesn't tell you anything useful, like how to fire it up (yes, I
> eventually figured out how to) or get it to do anything you might like to
> with a text editor, such as open a file. Totally worthless, unless you've
> been using it for the last 20 years.
>
> I suspect this language would get more attention if, say, there were a
> halfway decent development environment for it, or "how to get started"
> type files, such as exist for every other damned thing.
It has plenty. You're insisting on having an easy, Free, for
complete-newbies-who-don't-even-know-Emacs environment. On Linux.
No, we don't have that. You can get several Lisps on Windows or the
Mac that you start by clicking an icon. If you're experienced with
Unix and Emacs, getting the free lisps up on Unix isn't too hard. If
you use Debian, it's easy.
As for Hemlock, if you're willing to dismiss something before you get
to page 29 of the users manual, well, you'll probably get what you
deserve.
If you're actually looking for help, try making an honest effort, and
asking questions, not flaming.
--
/|_ .-----------------------.
,' .\ / | No to Imperialist war |
,--' _,' | Wage class war! |
/ / `-----------------------'
( -. |
| ) |
(`-. '--.)
`. )----'
As someone who has learned CL and used it extensively over the past
year, I agree with the OP. It's certainly possible to learn the tools
if one is persistent, and the community is often quite helpful, but
documentation often seems hard to find or even misleading or
incorrect.
As a simple example, it took me quite a while to get the
ILISP/hyperspec interaction working. The ILISP manual (p. 8)
indicates that the ilisp.emacs file provides sample instructions for
accessing the hyperspec (via Erik Naggum's package).
ilisp.emacs contains the following relevant section:
;;; Configuration of Erik Naggum's HyperSpec access package.
;; If you have a local copy of the HyperSpec, set its path here.
; (setq common-lisp-hyperspec-root
; "file:/home/joe/HyperSpec/")
; (setq common-lisp-hyperspec-symbol-table
; "/home/joe/HyperSpec/Data/Map_Sym.Txt")
;; Here's how to get the newest version of the CLHS:
;; <http://groups.google.com/groups?selm=sfwvgftux7g.fsf%40shell01.TheWorld.com>
Ignoring the fact that the groups.google link points to an old message
suggesting that one can get Version 5 of the Hyperspec by downloading
the Xanalys Personal Edition and unpacking the documentation (the
manual at least correctly indicates that Version 6 can be directly
downloaded from Xanalys with the correct URL), there is a problem with
this advice.
In the hyperspec, there is no file Map_Sym.Txt. There is a file
Map_Sym.txt, but not a Map_Sym.Txt. This error is not *that* easy to
track down (it doesn't just say "Map_Sym.Txt not found" for instance),
the fix for it is one keystroke to committed to CVS, and although I
reported the issue in March to the ilisp-help mailing list in March,
it hasn't been fixed (just checked today).
If this were totally isolated, that would be one thing, but this has
been fairly typical of my experience with CL tools. There's often
some little flub that takes a long time to track down, and when you
track it down, people say, "Oh, yeah, I remember that. You have to go
into the code and change this line. That's what I did when I had this
problem," which is better than nothing, but my experience with other
languages has been that the tools have a greater tendency to Just
Work.
rif
From: Lupo LeBoucher
Subject: Re: How does ILisp/HyperSpec work, and a general observation about Lisp culture
Date:
Message-ID: <DbqdnbejhddpfjeiRVn-tw@io.com>
In article <···············@five-percent-nation.mit.edu>,
rif <···@mit.edu> wrote:
>
>As a simple example, it took me quite a while to get the
>ILISP/hyperspec interaction working. The ILISP manual (p. 8)
>indicates that the ilisp.emacs file provides sample instructions for
>accessing the hyperspec (via Erik Naggum's package).
*snip*
>In the hyperspec, there is no file Map_Sym.Txt. There is a file
>Map_Sym.txt, but not a Map_Sym.Txt. This error is not *that* easy to
>track down (it doesn't just say "Map_Sym.Txt not found" for instance),
>the fix for it is one keystroke to committed to CVS, and although I
>reported the issue in March to the ilisp-help mailing list in March,
>it hasn't been fixed (just checked today).
Good god.
Thank you, sir, for pointing this out. It is indeed the problem I was
having. I should have been able to figure that out, based on the facts
that:
a) it worked on Windows (as I recall, in Bill's world, case doesn't
matter), but not on Unix and
b) it wasn't finding anything that is in the index, and Map_Sym.txt is the
index (yes, I had read enough of the code, code in a language I do not yet
grok, mind you, to know this).
But, silly me, I expected the documentation to be of slightly more than
marginal accuracy. Perhaps most people develop on Windows these days.
>If this were totally isolated, that would be one thing, but this has
>been fairly typical of my experience with CL tools. There's often
>some little flub that takes a long time to track down, and when you
>track it down, people say, "Oh, yeah, I remember that. You have to go
>into the code and change this line. That's what I did when I had this
>problem," which is better than nothing, but my experience with other
>languages has been that the tools have a greater tendency to Just
>Work.
That's all I'm saying.
I would like to see a larger CL community. I would have more neat
libraries to play with; ones that are of a more recent provenance than the
1980s, where I have a chance of talking to the guy who wrote them. I would
have more "recently learned" folk like you who could guide me through the
rough bits, such as exist in most software communities. Maybee by now,
CL would have been revised to include better packaging systems, and
more consistent interfaces to "alien" code and the OS (I think it is safe
at this point, to be rid of the idea that people are going to write OS in
Lisp). But, if the basic tools for getting started with CL are pestiferous
to use, that isn't ever going to happen. Who will bother to learn?
-Lupo
"To believe in one God you have to be a cretin. That is the only word for
it."-Michel Houellebecq. <··@io.com>
From: Daniel Barlow
Subject: Re: How does ILisp/HyperSpec work, and a general observation about Lisp culture
Date:
Message-ID: <87znf88eow.fsf@noetbook.telent.net>
··@io.com (Lupo LeBoucher) writes:
> I would like to see a larger CL community. I would have more neat
> libraries to play with; ones that are of a more recent provenance than the
> 1980s, where I have a chance of talking to the guy who wrote them. I would
And you believe that the best way to achieve this happy state is by
making highly negative "general observations" about the people who are
creating these tools and libraries?
> have more "recently learned" folk like you who could guide me through the
> rough bits, such as exist in most software communities. Maybee by now,
Oh, they exist. Unfortunately, when you start out by flaming people
you tend to find - regardless of the merits or otherwise of your case
- that they quite rapidly lose interest in helping you.
-dan
--
http://web.metacircles.com/cirCLe_CD - Free Software Lisp/Linux distro
From: Lupo LeBoucher
Subject: Re: How does ILisp/HyperSpec work, and a general observation about Lisp culture
Date:
Message-ID: <McWdnbhkPLEncDeiRVn-vw@io.com>
In article <···············@famine.OCF.Berkeley.EDU>,
Thomas F. Burdick <···@famine.OCF.Berkeley.EDU> wrote:
>··@io.com (Lupo LeBoucher) writes:
>
>> Yes, I have tried Hemlock; it was amusing in that, the documentation for
>> it doesn't tell you anything useful, like how to fire it up (yes, I
>> eventually figured out how to) or get it to do anything you might like to
>> with a text editor, such as open a file. Totally worthless, unless you've
>> been using it for the last 20 years.
>>
>> I suspect this language would get more attention if, say, there were a
>> halfway decent development environment for it, or "how to get started"
>> type files, such as exist for every other damned thing.
>
>It has plenty. You're insisting on having an easy, Free, for
>complete-newbies-who-don't-even-know-Emacs environment. On Linux.
I've been using x/emacs and Unix for the last 12 years, for heaven's sake!
>No, we don't have that. You can get several Lisps on Windows or the
>Mac that you start by clicking an icon. If you're experienced with
>Unix and Emacs, getting the free lisps up on Unix isn't too hard. If
>you use Debian, it's easy.
But none of this is the point. The languages themselves are quite easy to
install; almost as easy as, say, python. They appear to be robust and
pleasant interpretors to work with. There are even a half dozen of them to
choose from (which, in itself, is a bit bewildering to a novice).
The point is: the IDE doesn't work right out of the box. How, pray, am I
supposed to learn a language when one needs a more than casual
acquaintance of the language to make the IDE work? Bad enough that one
needs to download several parts, make them and glue them together to make
a complete IDE. Worse when they don't actually work.
>As for Hemlock, if you're willing to dismiss something before you get
>to page 29 of the users manual, well, you'll probably get what you
>deserve.
Brother Lupo assumes you're making a fun joke here.
Ha ha; fun.
If not: we are talking about an editor here. If I have to read 29 pages
into a malformed postscript file (incidentally, have you *looked* at the
documentation? I don't think you have -it looks like someone sneezed on
it -at least the version that comes with cmu-cl) in order to *start* the
thing, it is totally useless to a novice user. Incidentally, the
information that one needs to *start* the editor is on page 13, which is
absurd in itself, and, what is worse, the information given is *WRONG.*
Nowhere in the document can one find the correct information on how to
start the editor, viz:
(require: hemlock)
(ed)
In the end, I was able to figure this out by looking at dejanews. I found
the information ... posted in 1987.
That, sir, is the very definition of useless.
Sure, Hemlock will work for a guy who has used Hemlock since Jimmy Carter
was president, but it is no good for a novice user. No good at all.
-Lupo <··@pentagon.io.com>
"Java: the elegant simplicity of C++ and the blazing speed of Smalltalk"
From: Thomas F. Burdick
Subject: Re: How does ILisp/HyperSpec work, and a general observation about Lisp culture
Date:
Message-ID: <xcvislxvsnx.fsf@famine.OCF.Berkeley.EDU>
··@io.com (Lupo LeBoucher) writes:
> In article <···············@famine.OCF.Berkeley.EDU>,
> Thomas F. Burdick <···@famine.OCF.Berkeley.EDU> wrote:
>
> >No, we don't have that. You can get several Lisps on Windows or the
> >Mac that you start by clicking an icon. If you're experienced with
> >Unix and Emacs, getting the free lisps up on Unix isn't too hard. If
> >you use Debian, it's easy.
All of the above refer to getting a Lisp system *with* an IDE.
> But none of this is the point. The languages themselves are quite easy to
> install; almost as easy as, say, python. They appear to be robust and
> pleasant interpretors to work with. There are even a half dozen of them to
> choose from (which, in itself, is a bit bewildering to a novice).
>
> The point is: the IDE doesn't work right out of the box. How, pray, am I
> supposed to learn a language when one needs a more than casual
> acquaintance of the language to make the IDE work? Bad enough that one
> needs to download several parts, make them and glue them together to make
> a complete IDE. Worse when they don't actually work.
The commercial IDEs work, the free stuff works on Debian by just
installing packages, and there's a cookbook webpage for getting
started with ILISP
<http://cl-cookbook.sourceforge.net/emacs-ide.html>. ILISP is not
particularly user-friendly (though it's not the hardest Emacs package
I've had to configure), but the documentation situation is rapidly
improving. And again, the free trials of the commercial
implementations work just fine.
> >As for Hemlock, if you're willing to dismiss something before you get
> >to page 29 of the users manual, well, you'll probably get what you
> >deserve.
>
> Brother Lupo assumes you're making a fun joke here.
> Ha ha; fun.
>
> If not: we are talking about an editor here. If I have to read 29 pages
> into a malformed postscript file (incidentally, have you *looked* at the
> documentation? I don't think you have
Maybe there's something wrong with your postscript viewer. Anyway,
the version linked to from the Hemlock section of
<http://cmucl.cons.org/cmucl/> works fine. And yes, I'm a
semi-regular Hemlock user. I learned to use it by reading the manual:
first I looked at the table of contents, then I skimmed the
interesting-sounding parts. It was pretty easy.
> Nowhere in the document can one find the correct information on how to
> start the editor, viz:
> (require: hemlock)
(require :hemlock)
> (ed)
It's on the Hemlock page of the CMUCL website, clear as day.
> Sure, Hemlock will work for a guy who has used Hemlock since Jimmy Carter
> was president, but it is no good for a novice user. No good at all.
Well, I learned when Clinton was president, and it was pretty easy.
--
/|_ .-----------------------.
,' .\ / | No to Imperialist war |
,--' _,' | Wage class war! |
/ / `-----------------------'
( -. |
| ) |
(`-. '--.)
`. )----'
From: Lupo LeBoucher
Subject: Re: How does ILisp/HyperSpec work, and a general observation about Lisp culture
Date:
Message-ID: <kMqdnXwFsp--vTaiRVn-jg@io.com>
In article <···············@famine.OCF.Berkeley.EDU>,
Thomas F. Burdick <···@famine.OCF.Berkeley.EDU> wrote:
>··@io.com (Lupo LeBoucher) writes:
>
>> In article <···············@famine.OCF.Berkeley.EDU>,
>> Thomas F. Burdick <···@famine.OCF.Berkeley.EDU> wrote:
>
>> The point is: the IDE doesn't work right out of the box. How, pray, am I
>> supposed to learn a language when one needs a more than casual
>> acquaintance of the language to make the IDE work? Bad enough that one
>> needs to download several parts, make them and glue them together to make
>> a complete IDE. Worse when they don't actually work.
>
>The commercial IDEs work, the free stuff works on Debian by just
>installing packages, and there's a cookbook webpage for getting
>started with ILISP
><http://cl-cookbook.sourceforge.net/emacs-ide.html>.
Yes, I'm aware of that. Parts of it were very useful. The HyperSpec part
wasn't.
I am also aware of:
http://cl-cookbook.sourceforge.net/.emacs
Which didn't work for me either for reasons likely to remain mysterious.
> ILISP is not
>particularly user-friendly (though it's not the hardest Emacs package
>I've had to configure), but the documentation situation is rapidly
>improving.
I'm looking forward to that happy day.
That day has not arrived yet.
>> >As for Hemlock, if you're willing to dismiss something before you get
>> >to page 29 of the users manual, well, you'll probably get what you
>> >deserve.
>>
>> Brother Lupo assumes you're making a fun joke here.
>> Ha ha; fun.
>>
>> If not: we are talking about an editor here. If I have to read 29 pages
>> into a malformed postscript file (incidentally, have you *looked* at the
>> documentation? I don't think you have
>
>Maybe there's something wrong with your postscript viewer. Anyway,
>the version linked to from the Hemlock section of
><http://cmucl.cons.org/cmucl/> works fine. And yes, I'm a
>semi-regular Hemlock user. I learned to use it by reading the manual:
>first I looked at the table of contents, then I skimmed the
>interesting-sounding parts. It was pretty easy.
>
>> Nowhere in the document can one find the correct information on how to
>> start the editor, viz:
>> (require: hemlock)
> (require :hemlock)
Indeed; ooops.
I don't anticipate giving it much of a workout any time soon. I like xemacs
just fine, even if ilisp is kinda hackey.
>> (ed)
>
>It's on the Hemlock page of the CMUCL website, clear as day.
I will point out, again, that it ain't in the freakin' manual.
What's up with that, hmm?
>> Sure, Hemlock will work for a guy who has used Hemlock since Jimmy Carter
>> was president, but it is no good for a novice user. No good at all.
>
>Well, I learned when Clinton was president, and it was pretty easy.
Brother Lupo was busy learning other things during the Clinton era (like,
how to avoid being eaten by Berkeley homeless people on the way to Strada).
He hopes to make up for lost time without too many silly "the goddamned manual
is broken" or "you have to read 120 pages of screwey postscript to use this
editor" types of issues.
-Lupo
"Nothing shocks me. I'm a scientist." - Harrison Ford, as Indiana Jones
<··@io.com>
Lupo LeBoucher wrote:
> Here is what sucks: the HyperSpec thingee doesn't do anything. It claims
> 'cons' and '+' isn't a part of the HyperSpec, which I know to be untrue.
> Under other compilation conditions, it sends cryptic elisp debugger
> messages to the minibuffer thingee.
>
Perhaps it's not actually finding the hyperspec index?
On my system, I have extracted the hyperspec bundle in
/usr/local/lisp/CLHS6/, this contains a directory "HyperSpec" and two
.text files.
My ilisp-related emacs init .el file, lifted from the ilisp example,
contains (among other things)
;; If you have a local copy of the HyperSpec, set its path here.
(setq common-lisp-hyperspec-root
"file:/usr/local/lisp/CLHS6/HyperSpec/")
(setq common-lisp-hyperspec-symbol-table
"/usr/local/lisp/CLHS6/HyperSpec/Data/Map_Sym.txt")
And it works. One is a URL and one a unixoid filesystem path,
different things. That is what is given in the examples, but I know from my
personal mistake one can overlook the difference.
This may seem obvious too: you also have to have filesystem permissions to
read the hyperspec directory. You may need to chown and chmod things after
extracting the hyperspec.
The other thing which can trip you up is that sometimes a distro-packaged
Emacs may struggle to launch an external web browser, you may want to go to
the "Browse URL" customisation group in emacs and check that it is set up
sanely - some linux distros still have it set up to try to launch Netscape
4, which, of course, isn't even installed on most systems in this day and
age, though emacs is quite capable of spawning Konqueror, Moz or its own
internal w3.
Yet another thing that can trip you up, but it is rather unlikely by now, is
that if you are using an old ilisp and you have a shiny new hyperspec
downloaded from xanalys, there was a change in format of the hyperspec (I
don't recall the exact details, but it caused flamewars at the time) - just
means you need a newer ilisp package. But if you're using an even vaguely
up-to-date package, that's AFAIK unlikely.
More generally, a reason that getting started can be difficult is that a lot
of "setting up for common lisp development with the free stuff" advice
assumes prior familiarity with Emacs (as I kinda did above). Emacs itself
usually comes with its own online manual including tutorials, though, and
it is worth working through it.
······@mail.com (KanZen) writes:
> I've installed ILisp, copied ilisp.emacs to .emacs and tweaked the
> various paths within. But when I ruyn emacs and do "M-x cmulisp",
> visit a foo.lisp file and then try to evaluate (/ 1 0) I get the elisp
> stack trace, not the one from cmulisp.
Are you sure you didn't accidentally hit C-x C-e instead of C-c C-e
--
jan