[armedbear-devel] ASDF and ABCL
Dear ABCL developers, the one big thing missing in ABCL to fully pass the ASDF test suite is support for (directory "/usr/share/common-lisp/source/**/*.asd") Could you get it to work and update to the latest stable ASDF? That would be great! [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The trouble with opportunity is that it always comes disguised as hard work. — Herbert V. Prochnow
On 28 September 2010 00:33, Faré <fahree@gmail.com> wrote:
Dear ABCL developers, the one big thing missing in ABCL to fully pass the ASDF test suite is support for (directory "/usr/share/common-lisp/source/**/*.asd") Could you get it to work and update to the latest stable ASDF? That would be great!
I'll see what I can do. I implemented the normal wildcard support for DIRECTORY in abcl, I am familiar with the area.
On Sep 27, 2010, at 11:35 PM, Ville Voutilainen wrote:
On 28 September 2010 00:33, Faré <fahree@gmail.com> wrote:
Dear ABCL developers, the one big thing missing in ABCL to fully pass the ASDF test suite is support for (directory "/usr/share/common-lisp/source/**/*.asd") Could you get it to work and update to the latest stable ASDF? That would be great!
I'll see what I can do. I implemented the normal wildcard support for DIRECTORY in abcl, I am familiar with the area.
I'm still out of commission for contributing to ABCL for at least another two weeks due to "real-world" contraints, so won't be able to carry the torch at least until then. In surveying other Lisps for cribbing code ideas, SBCL was the only one that seemed to help (out of SBCL, clisp, and ECL that I looked at), but that was about as far as I got. -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."
On 30 September 2010 00:24, Mark Evenson <evenson@panix.com> wrote:
In surveying other Lisps for cribbing code ideas, SBCL was the only one that seemed to help (out of SBCL, clisp, and ECL that I looked at), but that was about as far as I got.
I wouldn't think it's ultimately all that complex to match 0-n directory levels when finding a :wild-inferiors.
On Sep 30, 2010, at 12:20 AM, Ville Voutilainen wrote:
On 30 September 2010 00:24, Mark Evenson <evenson@panix.com> wrote:
In surveying other Lisps for cribbing code ideas, SBCL was the only one that seemed to help (out of SBCL, clisp, and ECL that I looked at), but that was about as far as I got.
I wouldn't think it's ultimately all that complex to match 0-n directory levels when finding a :wild-inferiors.
I'm not claiming that it is that complex, only that if one wishes to see how another implementation handles such an algorithim in pure Lisp, then look at SBCL. -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."
On 30 September 2010 10:49, Mark Evenson <evenson@panix.com> wrote:
I'm not claiming that it is that complex, only that if one wishes
Oh, no offense intended.
to see how another implementation handles such an algorithim in pure Lisp, then look at SBCL.
Yeah, but that would spoil all the fun. :)
On 28 September 2010 00:35, Ville Voutilainen <ville.voutilainen@gmail.com> wrote:
On 28 September 2010 00:33, Faré <fahree@gmail.com> wrote:
Dear ABCL developers, the one big thing missing in ABCL to fully pass the ASDF test suite is support for (directory "/usr/share/common-lisp/source/**/*.asd") Could you get it to work and update to the latest stable ASDF? That would be great! I'll see what I can do. I implemented the normal wildcard support for DIRECTORY in abcl, I am familiar with the area.
Fixed by r12985 (http://trac.common-lisp.net/armedbear/changeset/12985), please test.
On 30 October 2010 11:09, Ville Voutilainen <ville.voutilainen@gmail.com> wrote:
On 28 September 2010 00:35, Ville Voutilainen <ville.voutilainen@gmail.com> wrote:
On 28 September 2010 00:33, Faré <fahree@gmail.com> wrote:
Dear ABCL developers, the one big thing missing in ABCL to fully pass the ASDF test suite is support for (directory "/usr/share/common-lisp/source/**/*.asd") Could you get it to work and update to the latest stable ASDF? That would be great! I'll see what I can do. I implemented the normal wildcard support for DIRECTORY in abcl, I am familiar with the area.
Fixed by r12985 (http://trac.common-lisp.net/armedbear/changeset/12985), please test.
Interestingly, ASDF 2.010, released yesterday, for performance reasons, has moved away from using (directory #p"**/*.asd") then filtering, towards computing subdirectories (which was much harder to do than should be, requiring 9 different variants for 11 implementations) and recursing manually, pruning as you go. It passes all the tests with ABCL. Please upgrade at your earliest convenience. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] You can tell whether a man is clever by his answers. You can tell whether a man is wise by his questions. — Naguib Mahfouz
On 10/30/10 8:51 PM, Faré wrote: […]
Interestingly, ASDF 2.010, released yesterday, for performance reasons, has moved away from using (directory #p"**/*.asd") then filtering, towards computing subdirectories (which was much harder to do than should be, requiring 9 different variants for 11 implementations) and recursing manually, pruning as you go. It passes all the tests with ABCL.
Please upgrade at your earliest convenience.
ABCL trunk upgraded to ASDF-2.010.1 in [r12986][1]. [1]: http://trac.common-lisp.net/armedbear/changeset/12986 -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."
On 31 October 2010 01:46, Mark Evenson <evenson@panix.com> wrote:
ABCL trunk upgraded to ASDF-2.010.1 in [r12986][1].
Thanks a lot! BTW, in the new numbering scheme I just inaugurated, 2.010 is stable, and 2.010.1 is a development version. In this case, 2.010.1 includes a GCL-specific patch, which won't affect you. I would recommend that in general, you should include a 2.0xx version in your official releases, not a 2.0xx.y version, unless the development version includes some ABCL-specific fix. This will help keep things more stable and restrict the number of different versions of ASDF found in the wild. No need to downgrade to 2.010, though. Just keep that in mind next time you upgrade. Or maybe you think this is a bad move, and have better ideas as to how ASDF versions may be numbered. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Wealth, like happiness, is never attained when sought after directly. It comes as a by-product of providing a useful service. — Henry Ford
On 10/31/10 9:58 AM, Faré wrote:
On 31 October 2010 01:46, Mark Evenson<evenson@panix.com> wrote:
ABCL trunk upgraded to ASDF-2.010.1 in [r12986][1].
Thanks a lot!
BTW, in the new numbering scheme I just inaugurated, 2.010 is stable, and 2.010.1 is a development version. In this case, 2.010.1 includes a GCL-specific patch, which won't affect you. I would recommend that in general, you should include a 2.0xx version in your official releases, not a 2.0xx.y version, unless the development version includes some ABCL-specific fix. This will help keep things more stable and restrict the number of different versions of ASDF found in the wild.
No need to downgrade to 2.010, though. Just keep that in mind next time you upgrade.
Or maybe you think this is a bad move, and have better ideas as to how ASDF versions may be numbered.
Thanks for the heads up on the versioning change. I must confess that I don't understand git very well, being more of a Mercurial partisan, so I had just issued unix$ git fetch origin unix$ git pull which had worked before to follow the stable branch of ASDF. I don't expect that you are mixing development on the stable branch, so I'll have to poke around the git documentation to figure out what I am doing wrong. -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."
On 31 October 2010 02:10, Mark Evenson <evenson@panix.com> wrote:
Thanks for the heads up on the versioning change. I must confess that I don't understand git very well, being more of a Mercurial partisan, so I had just issued
unix$ git fetch origin unix$ git pull
which had worked before to follow the stable branch of ASDF.
Oh, but the default branch, master, is where development is currently ongoing. Release happens in the "release" branch: git checkout release # maybe with option -b the first time. git pull origin/release
I don't expect that you are mixing development on the stable branch, so I'll have to poke around the git documentation to figure out what I am doing wrong.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Our love was a supernova, bursting with an immense light only to wane quickly, leaving behind a black hole.
On 31 October 2010 11:10, Mark Evenson <evenson@panix.com> wrote:
Thanks for the heads up on the versioning change. I must confess that I don't understand git very well, being more of a Mercurial partisan, so I had just issued unix$ git fetch origin unix$ git pull which had worked before to follow the stable branch of ASDF.
I don't expect that you are mixing development on the stable branch, so I'll have to poke around the git documentation to figure out what I am doing wrong.
You can use a certain version by doing a git checkout after the fetch. The pull actually does a fetch and rebase (merge), you don't need to fetch and pull separately. Anyway, checkout uses a certain version. If there are branches, you can checkout a given branch.
participants (3)
-
Faré -
Mark Evenson -
Ville Voutilainen