[asdf-devel] clisp and asdf

[Faré]

On Mon, Jan 21, 2013 at 11:51 AM, Sam Steingold <sds at gnu.org> wrote:
>> * Faré <snuerr at tznvy.pbz> [2013-01-20 22:46:35 -0500]:
>>
>> Of all the lisp implementations, CLISP has given me the most trouble
>> with ASDF upgrade.  Unlike all other Lisps, it won't let me undefine
>> functions and redefine them in the same fasl file.  Somehow, I always
>> end up with CLISP either complaining that a function signature doesn't
>> match (was or wasn't a generic function with an incompatible
>> lambda-list), because I couldn't undefine it, or that it is undefined
>> and never redefined, because I undefined it and CLISP throws away my
>> redefinition somehow. I tried to selectively unintern symbols for
>> functions thus upgraded, but that also fails. Yet all these things
>> work in about all other Lisps.
>>
>> In the end, I punted: I just rename the old ASDF package to
>> ASDF-${version} early on, but that messes with any package that :use's
>> ASDF, and then CLISP will complain again, unless ASDF is the very
>> first thing upgraded before anything uses it.
>>
>> Could CLISP be made more upgrade-friendly?
>
> When you actually do change function signatures, a warning is, IMO,
> quite appropriate.  I am not really sure why you want to silence them.
>
A warning when the signature changes is quite OK.
My problem is the unability to override the previous signature.
If I fmakunbound the function then redefine it in the same file,
it looks like the new definition never appears.

I will try to extract some reduced example.

>> (another thing I don't understand that may or may not be related is
>> the .lib output files produced together with a fasl,
>> when and how they are used, and what I should or should not do with them.)
>
> http://clisp.org/impnotes/require.html#lib-files
>
Oh, that's cool. Does the .lib also include the user-supplied
(EVAL-WHEN (:COMPILE-TOPLEVEL) ...) side-effects?
If so, that's great -- it's like the CFASLs of SBCL
(except it probably pre-dates them).

—♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org
Die Mathematiker sind eine Art Franzosen: redet man zu ihnen, so übersetzen
sie es in ihre Sprache, und dann ist es alsobald ganz etwas anderes.
[Les mathématiciens sont comme les Français: quoiqu'on leur dise,
ils le traduisent dans leur propre langue, et cela devient alors
quelque chose de complètement différent.] — Johann Wolfgang von Goethe