Re: [Cclan-commits] CVS: asdf asdf.lisp,1.81,1.82

On Wed, 05 May 2004 17:22:25 +0100, Christophe Rhodes <[email protected]> wrote:

> Erm, say what?  Corman Lisp doesn't have a file compiler?  Or maybe
> it has a broken COMPILE-FILE-PATHNAME?

Neither one of these - see my recent mail to the CCLAN list.

> In any case, this looks like a bad way of solving this problem.  What
> happens when Corman's bug is fixed?  Some asdfs will work, others
> won't; confusion will ensue (I draw the gentle reader's attention to
> <http://www-jcsu.jesus.cam.ac.uk/~csr21/papers/features.pdf>).

I've read the paper but I see no easy way to apply it to the problem
at hand - see below.

> Preferred fixes, in order:
>   * put pressure on Roger Corman to implement compile-file-pathname
>     (or whatever is missing) in Corman Lisp asap, and simply ignore
>     broken versions;

That's unlikely to happen. Roger doesn't seem to work on Corman Lisp
on a regular basis. Looks like he has to pay his rent and I can't
blame him for that.

And if he does there are /lots/ of things to fix in CCL and he's aware
of them.

>   * add
>       (eval-when (:compile-toplevel :execute)
>         (unless (ignore-errors (compile-file-pathname "foo.lsp"))
>           (error "This Lisp is too far from Common Lisp, sorry")))
>     at toplevel near the top of the file;

This test won't work because CCL /has/ COMPILE-FILE-PATHNAME.

>   * add
>       (eval-when (:compile-toplevel :execute)
>         (unless (ignore-errors (compile-file-pathname "foo.lsp"))
>           (pushnew :broken-compile-file-pathname *features*)))
>       #-broken-compile-file-pathname
>       (eval-when (:load-toplevel)
>         (unless (ignore-errors (compile-file-pathname "foo.lsp"))
>           (error "Sorry, you have a broken COMPILE-FILE-PATHNAME")))
>     at toplevel near the top of the file, and then use
>     BROKEN-COMPILE-FILE-PATHNAME rather than CORMANLISP as the feature
>     tester.

Same here. The problem is the FASL loader and it /sometimes/ works and
sometimes doesn't. I'd be able to build a test to create a
:CCL-BROKEN-FASL-LOADER feature but it would be a lot more complicated
and quite longish.

The current situation is that ASDF doesn't work with CCL (or rather
vice versa) and I was trying to fix this. This involves a couple of
patches to CCL itself (which I'm done with) and these two small
patches to ASDF. The next step would be a short website which explains
to CCL users what to do to be able to use ASDF with Corman Lisp.

Once the official Corman distribution is fixed enough to use ASDF out
of the box we can think about ways to make ASDF distinguish between
"old" and "new" versions of CCL.

However, if the CCLAN policy is not to support any Lisp that's not
100% ANSI compliant then feel free to remove the patches. And don't
forget the stuff that's only in there for CMUCL, too... :)

Cheers,
Edi.


-------------------------------------------------------
This SF.Net email is sponsored by Sleepycat Software
Learn developer strategies Cisco, Motorola, Ericsson & Lucent use to 
deliver higher performing products faster, at low TCO.
http://www.sleepycat.com/telcomwpreg.php?From=osdnemail3