Re: Re: [Cclan-commits] CVS: asdf asdf.lisp,1.89,1.90

Edi Weitz <[email protected]> writes:

>> I claim that the use of *features* for functionality testing is
>> almost always wrong, in that it does not have a fine enough
>> granularity; it's usually better to test for the functionality you
>> actually need.
>
> I agree in theory.  In practise I think it's nice to have this as a
> convenience feature.  One could also argue that those who agree with
> your POV don't have to use these features - they don't really hurt.
> With Alan's patch one has the choice to use them or not.  Without
> Alan's patch one is (well, in a way) forced to see things your way.  I
> vote for more choice... :)

Heh, well, there's choice and choice... this kind of thing makes it
potentially harder to write asdf-to-vendor-defsystem converters, as
the behaviour needs to be emulated -- essentially, you write into your
software a dependency on its build tool (which to me seems pretty
weird.)

In order of my preference, then, I'll

* revert this patch completely, or
* push the system name onto *MODULES* instead, or
* make the pushed symbol asdf::systemname rather than
  :asdf-systemname, or
* sulk quietly in a corner.

Alan?

Cheers,

Christophe


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click