[armedbear-devel] ASDF 2.013 released

Faré fahree at gmail.com
Sun Mar 20 16:08:58 UTC 2011


> We were at asdf-2.012.  I just [committed asdf-2.013][1] for you (and us
> all, really).
>
> [1]: http://trac.common-lisp.net/armedbear/changeset/13253
>
Thanks a lot! I already have a patch 2.013.1, but it only matters to
people upgrading from 2.008 and earlier, which doesn't concern you.

> ASDF is working just fine with ABCL as far as we know.  We don't currently
> have any implementation specific patches.
>
I didn't suppose you had any. Just boilerplate.
This message doesn't represent the views of my employer, nor of my window.

> Since we are a JVM-based Lisp, the one request we seem to get from our user
> community would be the ability to distribute ASDF systems with the FASLs (in
> our case Java bytecode plus some glue to be evaled) already present, i.e. to
> skip the compile step.  I know that ECL was nosing for something similar to
> support deployment to systems without gcc present, but I didn't really see
> this result in a solution (or maybe I missed it).  Is there a
> semi-officially blessed way to distribute ASDF systems with FASLs while
> skipping the compilation step and somehow bypassing the centralized FASL
> repository on a per-system basis?
>
Not at this moment. I'm not working on it, and know no one who is,
but I'd merge code if someone provided it.

That said, you can already distribute precompiled code
that includes both source and its precompiled binaries,
and just setup your source registry and output-translations
to point to the proper locations before you load-system it.
To remove the source, you could also cheat and replace all files
by empty files with a properly old timestamp
(either copied from the original or set to the epoch).

Alternatively, you could get a list of all the object files
by a clever use of TRAVERSE, then package them into a library or archive,
and have a non-ASDF way of loading said library or archive.

> We're currently at abcl-0.25.0.  Any idea how to get us into the Debian
> packaging stream?  Is there someone who specializes in maintaining CL
> related packages that we could lobby?
>
You could contact the <cl-debian at common-lisp.net> mailing list.
Considering how undermanned they are, what they will probably offer you is
to have you upload packages to mentors.debian.net and bless your packages
when you notify them. That's how I do it myself for ASDF (see the debian/
subdirectory in git). It might be slightly more involved for you since
you're compiling stuff, but I'm pretty sure there's a java helper taking
most of the hassle out of packaging java applications. The main maintenance
item when you upload a new version is updating the changelog and rerunning
a package-and-release script.

> If anyone reads this who would like help with packaging ABCL for their
> distribution system, we would be happy to work to make such integration as
> painless as possible.  Just drop us a post here on armedbear-develop, and we
> should be able to followup fairly quickly.


>> PS: I created a mailing-list asdf-announce.
>> Maybe someone amongst you could join it,
>> so I don't need to push my announces to this mailing-list.
>> Or maybe I can just Cc: my announce to this mailing-list
>> and not bother with further explanations?
>
> You are welcome to cc: your announce here to armedbear-devel and skip the
> personalized request if that saves you time.  I'm pretty tuned into the CL
> logosphere via #lisp, planetlisp, and c.l.l so I usually ambiently know that
> you're making a release.
>
I admit I don't frequent c.l.l anymore. Should I be posting announces there?

> Thanks for the excellent support with ASDF!  Although Quicklisp gets all the
> headlines, I know that without the minimum level of sanity introduced with
> ASDF2 and its subsequent maintenance releases, such contemporary CL
> distribution mechanisms would have been nearly impossible.
>
Don't underestimate how much crap Zach could work around.
But ASDF2 is still useful in simplifying Quicklisp a bit,
and bringing sanity to non-Quicklisp users, too.

[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
Nothing in the world can take the place of persistence.
Talent will not; nothing is more common than unsuccessful men with talent.
Genius will not; unrewarded genius is almost a proverb.
Education will not; the world is full of educated derelicts.
Persistence and determination alone are omnipotent.
The slogan 'Press On' has solved and always will solve
the problems of the human race.
	— Calvin Coolidge




More information about the armedbear-devel mailing list