Re: resolution of symlinks in asdf

Christophe Rhodes <[email protected]> wrote:

>> Kevin Layer <[email protected]> writes:
>> 
>> > cliki.net appears to be done, and has ever since I've been trying to
>> > read about ASDF-Binary-Location.  My remaining action on this matter
>> > is to read about that.
>> 
>> Have you considered and rejected the addition to asdf systems of the
>> unresolved system pathname?  If this wouldn't provide a solution to
>> your problem, it would be good to know why.
>> 
>> To make things a little more concrete, I think that (in the presence
>> of that construct, which I've called system-unresolved-pathname below)
>> the relevant output-files method would look like (untested):

I believe it would work.  I thought of it, too, but gave it lower
ranking in the solution space because the output-files method will
have to be replicated in each .asd file, since there are 14 .asd files
that make up this new product.  I never like duplicating code, for the
obvious reasons; recall, we don't have site-specific init files, as a
matter of policy.

I haven't yet looked at the asdf-binary-locations.tgz that Robert sent
me, and probably won't until tomorrow (if I'm lucky).

>>   (defmethod output-files :around ((o compile-op) (c component))
>>     (let ((results (call-next-method)))
>>       (let ((binary-path-root (system-unresolved-pathname (component-system c)))
>>             (default-path-root (system-pathname (component-system c))))
>>         (mapcar 
>>          (lambda (path)
>>            (merge-pathnames (enough-namestring path default-path-root)
>>                             binary-path-root))
>>          results))))
>> 
>> Cheers,
>> 
>> Christophe

Kevin