On Aug 16, 2005, at 3:49 PM, Paolo Amoroso wrote:
> Marco Antoniotti <marcoxa@...> writes:
>> On Aug 16, 2005, at 3:30 PM, Paolo Amoroso wrote:
>>> Sam Steingold <sds@...> writes:
>>>> is it true that "defsystem is dead and everyone should switch to
>>> I don't know whether DEFSYSTEM is dead. But I would use CLOCC
>>> subsystems much more if there was support for ASDF.
>> But MK:DEFSYSTEM supports subsystems in a nicer way than ASDF. So,
>> the question is why don't you give it a spin.
> I probably didn't explain correctly what I meant. By "CLOCC
> subsystem" I simply mean some library or tool included in CLOCC. I
> would like to be able to to with ASDF do something like:
> (asdf:operate 'asdf:load-op :cllib) ; src/cllib
> or even:
> (asdf:operate 'asdf:load-op :gnuplot) ; src/cllib/gnuplot
> or also:
> (defsystem my-app
> :depends-on (net shell gnuplot) ; src/port/net, src/port/shell,
On that I agree. CLOCC does feel a bit "monolithic".
> Is this what you mean?
MK:DEFSYSTEMs are ASDF-INSTALLABLE for one thing. Doing (REQUIRE
"some-mk-defsystem") on CMUCL works as expected, and above all, the
setup of MK:DEFSYSTEM is now much easier.
Essentially you just need to tell MK:DEFSYSTEM where your packages are,
without having to link in the .system file (provided that the package
unpacks in a directory contaning a .system file with the same name).
Moreover, now "component" systems work intuitively and can be made
dependent on other components as well. This considerably facilitates
the construction of hierarchical systems (witness the struggle I just
went through to make CLSQL build under Windows).
> In any case, I personally feel more
> "comfortable" with ASDF, and use it for all my projects.
Of course. I have no problems with that. I just want people to know
that MK:DEFSYSTEM is live and well. :)
Marco Antoniotti http://bioinformatics.nyu.edu
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
715 Broadway 10th FL fax. +1 - 212 - 998 3484
New York, NY, 10003, U.S.A.