I'd like to hear (er, read) what you think about it and how it can be improved.
Finally, I'm willing to temporarily volunteer as a "porting maintainer", so if anybody is willing to add support for other Lisps or make the whole shebang work on MS Windows please send patches to my email address.
Edi Weitz <e...@agharta.de> writes: > I have ported Daniel Barlow's ASDF-INSTALL[1] to CMUCL, CLISP, > AllegroCL, and LispWorks.
I assume that if you have ASDF-INSTALL running on CLISP, then ASDF must also work on CLISP. I checked out ASDF from CVS and tried to load asdf.lisp using CLISP 2.32, and can't get past this error pertaining to most of the `formatter' calls involving format strings with the "~@<...~@:>" pattern:
,----[ CLISP loading asdf.lisp ] | [1]> (load "asdf") | ;; Loading file /usr/local/lib/asdf/asdf.lisp ... | *** - Error: ~:@> not implemented | Current point in control string: | ~@<erred while invoking ~A on ~A~@:> | | | Break 1 ASDF[2]> `----
As these `formatter' calls are pervasive throughout ASDF, it's hard to see how CLISP users are using ASDF. Should I take this discussion to cclan-l...@lists.sf.net, or can we carry on here?
>> I have ported Daniel Barlow's ASDF-INSTALL[1] to CMUCL, CLISP, >> AllegroCL, and LispWorks.
> I assume that if you have ASDF-INSTALL running on CLISP, then ASDF > must also work on CLISP. I checked out ASDF from CVS and tried to load > asdf.lisp using CLISP 2.32, and can't get past this error pertaining > to most of the `formatter' calls involving format strings with the > "~@<...~@:>" pattern:
It's not actually directly related to FORMATTER; see below.
> ,----[ CLISP loading asdf.lisp ] > | [1]> (load "asdf") > | ;; Loading file /usr/local/lib/asdf/asdf.lisp ... > | *** - Error: ~:@> not implemented > | Current point in control string: > | ~@<erred while invoking ~A on ~A~@:> > | | > | Break 1 ASDF[2]> > `----
> As these `formatter' calls are pervasive throughout ASDF, it's hard to > see how CLISP users are using ASDF. Should I take this discussion to > cclan-l...@lists.sf.net, or can we carry on here?
I inserted the FORMATTER calls as a workaround to clisp's failure to implement ~< ~:> in regular FORMAT in released clisp versions prior to 2.32; the output from FORMATTER was not correct, but it was acceptable.
Of course, with the new earlier error reporting, asdf does not work at all on clisp-2.32. So the FORMATTER calls now just obscure the issue, which is that clisp does not implement the Common Lisp standard. However, the failure mode for clisp-2.31 and earlier was significantly less fatal to asdf's ambition to run even on slightly non-conforming implementations.
I suppose the good news is that now clisp is totally broken in this respect I can remove the FORMATTERs again :-)
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)
>> I have ported Daniel Barlow's ASDF-INSTALL[1] to CMUCL, CLISP, >> AllegroCL, and LispWorks.
> I assume that if you have ASDF-INSTALL running on CLISP, then ASDF > must also work on CLISP. I checked out ASDF from CVS and tried to > load asdf.lisp using CLISP 2.32, and can't get past this error > pertaining to most of the `formatter' calls involving format strings > with the "~@<...~@:>" pattern:
Yes, sorry I forgot about that one. I tried to compile ASDF when I installed 2.32 shortly after its release and it didn't work - same problem. It turned out that I was able to just load the .fas file that I created with 2.31. Maybe I should provide this .fas file as a workaround until CLISP can compile ASDF.
Edi Weitz <e...@agharta.de> writes: > I tried to compile ASDF when I installed 2.32 shortly after its > release and it didn't work - same problem.
So is this a regressive bug in 2.32?
> It turned out that I was able to just load the .fas file that I > created with 2.31. Maybe I should provide this .fas file as a > workaround until CLISP can compile ASDF.
That would be helpful. I'd like to start using ASDF, so even a temporary workaround would be sufficient.
Is this omission a known problem with CLISP 2.32? Do we need to take up discussion the CLISP mailing list?
>> It turned out that I was able to just load the .fas file that I >> created with 2.31. Maybe I should provide this .fas file as a >> workaround until CLISP can compile ASDF.
> That would be helpful. I'd like to start using ASDF, so even a > temporary workaround would be sufficient.
The file is available from <http://weitz.de/files/asdf.fas> - I've also updated the docs. IIRC CLISP's FAS files are platform-neutral. I hope I'm right. (Mine is from a Linux x86 system compiled with - I think - 2.31.)
> Is this omission a known problem with CLISP 2.32? Do we need to take > up discussion the CLISP mailing list?
I'm not following CLISP's development very closely but from what I remember and from what Christophe has posted it seems like not all of FORMAT/PPRINT is currently implemented and 2.32 raises an error where earlier versions quietly did the wrong thing. It may even be that I'm the guilty party - see
Edi Weitz <e...@agharta.de> writes: > The file is available from <http://weitz.de/files/asdf.fas> - I've > also updated the docs. IIRC CLISP's FAS files are platform-neutral. I > hope I'm right.
Well, it loads fine on Cygwin/Windows 2000 with CLISP 2.32. Thanks. I'll actually try using it later today and report back.
> Looks like a lack of developers wishing to fix this.
On Tue, 13 Jan 2004 20:49:30 +0100 Edi Weitz wrote:
> The file is available from <http://weitz.de/files/asdf.fas> - I've > also updated the docs. IIRC CLISP's FAS files are platform-neutral. I > hope I'm right. (Mine is from a Linux x86 system compiled with - I > think - 2.31.)
I've now added some Windows support for Allegro CL and LispWorks. (The code for CLISP is also there but it has network-related problems. Maybe someone else wants to debug this.) Make sure to download the newest version before you try it.
