Re: [clisp-list] asdf 2

[Faré <fa...@gm...>]

While I'm at it - ABCL, CCL, CMUCL, ECL, SBCL all have a hook whereby
cl:require will attempt loading modules via ASDF. It would be nice if
CLISP offered the same facility.

[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
Politicians are like diapers: they must be changed often.
And for the same reasons. [Also, adults don't need either of them. — Faré]




On 3 August 2010 18:40, Faré <fa...@gm...> wrote:
> Dear Sam,
>
> thanks for the quick answer.
>
> On 3 August 2010 13:04, Sam Steingold <sd...@gn...> wrote:
>>> Should I send a patch that consists in creating a new subdirectory of
>>> modules? Would you accept it? If so, which existing module may I
>>> cargo-cult, considering that ASDF is one single lisp file?
>>
>> Oh, you want _us_ to ship asdf...
> Yes, that would be great.
>
>> do people want that?
> With asdf2, unlike asdf1, you can easily upgrade from some asdf to
> another one, so if someone isn't fully satisfied with the
> clisp-provided asdf, he can still use it to portably upgrade to his
> installed asdf as easily as:
> (require :asdf)
> (asdf:load-system :asdf)
> With asdf1, you were kind of stuck, and indeed the case for including
> yet another divergent version of asdf with the implementation was less
> compelling.
>
>> if yes, should it be a base module?
>> (http://clisp.cons.org/impnotes/modules.html#base-modules)
> I think it should be indeed.
>
>> do other lisps come with asdf ootb?
>> (I know sbcl does, how about gcl, ecl, abcl &c)
> At least ABCL, CCL, CMUCL, ECL, SBCL include asdf2 ootb.
> ASDF2 unlike ASDF1 now works with GCL 2.7 (but doesn't pass all
> regression tests),
> and I should indeed contact the authors for inclusion, too.
>
>> PS. a quick look at #+clisp on asdf.lisp reveals the following:
>> getenv: I suggest (import 'ext:genenv) instead of (defun getenv)
> I don't want to import in some implementations, not in others. For
> portability reasons, I prefer a symbol with uniform semantics
> everywhere, including symbol package and bound function semantics. Too
> bad if you can't (setf (asdf-utilities:getenv "FOO") "bar") on clisp -
> you shouldn't be relying on it portably, anyway.
>
>> get-uid: I suggest (setf fdefinition) instead of defun
> Same here.
>
>> probe-file*: I suggest probe-pathname instead of (ignore-errors (truename
>> p))
> Oh, thanks a lot.
>
> Base on that feedback, I committed asdf 2.113. Can you test it? If it
> works for you, I'll release it as 2.005 so you may include the latest
> stable release to clisp. (Though I'll also offer the GCL guys to
> provide feedback.)
>
> [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
> None are more hopelessly enslaved than those who falsely believe they are free.
>        — Johann Wolfgang von Goethe
>