Re: ASDF bug?

Larry Hunter wrote:
>    (> (apply #'min (mapcar #'file-write-date out-files))
>       (apply #'max (mapcar #'file-write-date in-files)))
>  
> the spec allows file-write-date to return NIL, which causes max (in
> this case) to complain about a non-number.  This should probably be
> patched to something like
> 
>   (mapcar (lambda (file) (or (file-write-date file) 0)))

That'd be reasonable. Perhaps that'd also a location where it'd be
appropriate to signal an error that an requisite file is missing. I'm
not familiar enough with the execution path to know if a missing file
in that function is always a missing file error.

> First, I think ".cl" ".lisp" and ".lsp" are all pretty common source
> file types, and should be accomodated somehow.  Second, it's a pretty
> opaque error message, and there is no indication in the documentation
> that lisp source files must have type "lisp".  

I agree those are common source file types. I won't speak for ASDF's
author, but in the past my understanding is that rather than
attempting to make stock ASDF handle common senarios, Dan prefers to
support extensions to override stock behavior.

The commonly quoted method to handle non ".lisp" cl source file types
with ASDF is:

(defmethod source-file-type ((c cl-source-file)
                             (s (eql (find-system 'foo))))
  "cl")

If Dan decides to support multiple file types, it ease the library
packaging I do for Debian so that I could drop the above kind of
source-file-type method specializations.

Best,

Kevin


-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click