- armedbear-devel - mailman3.common-lisp.net

[armedbear-devel] Branching 0.16.x
by Erik Huelsmann 04 Sep '09

04 Sep '09

With the recent changes to our Ant build to create ZIP files with CRLF line endings and LF line endings in the TAR archive, I think we have enough basis to create a branch for release management of our 0.16.x release line. Should further fixes be required, they can be backported. Creating the release branch now frees up the trunk for further development. Are there any last minute fixes that need to be applied? From the silence on the list this last week I concluded there probably are not. I hope to be able to create a release before the end of the weekend, however, no guarantees! Bye, Erik.

2 1

[armedbear-devel] eol-style problem with abcl-src-0.15.tar.gz
by Mark Evenson 01 Sep '09

01 Sep '09

Somehow we seemed to have shipped [the abcl-0.15.0 source][1] with a CR+LF file-ending for 'abcl.in', which means when built under a 'NIX one gets the slightly mysterious message "#!/bin/sh -- interpreter not found. It is slightly surprising that we haven't gotten more complaints on this (or maybe everyone builds from SVN?). The svn:eol-style is now set to 'LF' in SVN for 'abcl.in', which does the "right thing" when building the source tar balls, so abcl-0.16 should be ok, as long as whoever builds the release doesn't somehow override this is local SVN settings Looking through the SVN log for metadata changes to 'abcl.in' doesn't seem to report anything. Maybe there is somehow to report metadata changes, but I don't see how to do this quickly. We have the currently unused ability to fix line endings in the relevant Ant targets. Maybe I should revisit them? If so, what would be the correct policy? [1]: http://downloads.sourceforge.net/armedbear-j/abc/abcl-src-0.15.0.tgz -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."

2 4

[armedbear-devel] Release planning: two month intervals, but which months?
by Erik Huelsmann 31 Aug '09

31 Aug '09

Last year we set a bi-monthly release planning. Last spring, we released once or twice within 6 weeks. This summer it took me - due to holidays - more time to get our trunk into a releasable state. Or so it must have looked from the outside. What I should have communicated is that I'd like to set our bi-monthly release planning to release around the first weekend of the odd-numbered months. Why? Well, it allows the release schedule to continue without suffering from most of the holidays: summer holidays (at least mine) are mostly during august; Christmas isn't around a first weekend and October holidays are rarely close to the first weekend of November. This means that February holidays may interfere with our release planning, but if I look at school holidays, these are mostly in the middle or the end of February. This plan would also mean that a release should be cut around next weekend. Then there's the issue of me being the only one with the right access to SourceForge to upload ABCL release files. We can do two things about that: 1. Give others upload access to SourceForge 2. Host our downloads on c-l.net The second option would make any committer a release manager, if we decided to do that [because we would commit the releases to our repository, which automatically updates our site]. The reason our downloads are currently hosted on SF is that they provide nice download stats. If we decide we don't need those or that we can create those stats ourselves on c-l.net, there's no reason to continue with SF. What are your comments? Bye, Erik.

2 2

[armedbear-devel] BLOCK-NODE -> specific node types done; let's go for a release
by Erik Huelsmann 31 Aug '09

31 Aug '09

Hi everybody, Over the past weeks, I've been rewriting our generic use of BLOCK-NODE structures to specific structures for every node type. There are TAGBODY-NODEs, CATCH-NODEs, etc. Although I introduced that change for a reason: I want to be able to analyse the nesting set up by pass1 from outer to inner blocks. This gives us some opportunities in preventing functions from becoming closures. Currently every function which is enclosed in a block which defines a closure, becomes a closure. That's too generic and doesn't need to be. Well, as I said, even though I want to introduce changes to that extent, I'm done changing the existing structure. This means I think we're in shape to go for a release. Everybody please run your test suite and your own applications to see if we still meet the quality levels we've delivered in the past releases (or, preferably: better!) I've updated changes, but there are probably a lot more tasks to be addressed. Just to name one: did 'build.xml' get fixed for releasing with Java 1.5? Or should I use 1.6 next time? Bye, Erik.

3 5

[armedbear-devel] BUG: problem with THE type checks in compiler
by eric.marsden＠free.fr 31 Aug '09

31 Aug '09

Hi, There is a problem in the recent improvements to the way the compiler handles THE type checking. File-compiling the code below leads to an error "The value 6 is not of type FOO". (defconstant +foo-OK+ (the (integer 0 10) 5)) (deftype foo () '(integer 0 10)) (defmacro assemble-foo (x) `(the foo ,x)) (defconstant +foo+ (assemble-foo 6)) Eric

3 4

[armedbear-devel] No JSR-223 code in abcl-src-0.15.tar.gz
by Alessio Stalla 28 Aug '09

28 Aug '09

Greetings all, I have downloaded the 0.15 release of abcl from the official site, to use it in a side-project at work. I found out that the archive does not contain the src/org/armedbear/lisp/scripting directory, where the JSR-223 stuff resides. Worse, the archive does contain the META-INF/services/javax.script.ScriptEngineFactory, so when built the jar actually declares it contains a JSR-223 implementation, but it does not. How is the archive built? I think I saw tasks in the Ant-based build file to zip the source, maybe those don't include the scripting/ directory. If that is the case, I can fix it for build.xml. I don't know how the Lisp based build works, though. Alessio

2 3

[armedbear-devel] Embedding ABCL into IntelliJ IDEA and writing plugins
by Anton Vodonosov 23 Aug '09

23 Aug '09

Hello, I created IntelliJ IDEA plugin that embeds ABCL, allows to connect from SLIME and extend IDEA using Common Lisp. I haven't contrived any useful functionality (that is easy to implement), therefore, for the sake of example, the plugin ads to IDEA commands yank and yank-pop, like in Emacs. Source may be found here: http://github.com/avodonosov/abcl-idea/tree/master Binary distro: http://cloud.github.com/downloads/avodonosov/abcl-idea/abcl-idea.zip There is a readme, check it for more details. Some conclusions from my experience: - it is absolutely possible to create IDEA plugins using ABCL. - Jfli interface to java is currently more convenient than native ABCL's - ABCL's performance of slime fuzzy-completion is still unsatisfying (it is the only little unsatisfying thing for me) - ABCL is more suited to be embedded into J editor, than into something else. For example Interpreter.kill method stops ABCL like: if (jlisp) { ... close input output streams } else System.exit(status); I.e. if we call Interpretter.kill() in IDEA plugin, it will exit whose IDEA process This is not a big problem, because with having Interpreter source code we may But having source code we can copy/past desirable kill behavior - closing streams. But it would make sense to reactor it a little to be more embedding friendly. If somebody is interested to write IDEA plugins in Lisp, my plugin may be used as a basis for further extension. Best regards, - Anton

3 4

Re: [armedbear-devel] progress on Maxima + ABCL -- light at the end of the tunnel!
by Erik Huelsmann 08 Aug '09

08 Aug '09

On Thu, Jul 30, 2009 at 8:04 PM, Robert Dodier<robert_dodier(a)yahoo.com> wrote: > > --- On Thu, 7/30/09, Erik Huelsmann <ehuels(a)gmail.com> wrote: > >> Another question: I'm running these profiles too, of >> course. On my system rtest_integrate takes forever to run, >> could that be this is the main cause it takes that long? >> Should we start profiling only that test set? > > Yes, that sounds reasonable. The integration code makes > use of the sign-inference code, to which DGR and DLS belong, > so maybe it is actually true that they are doing something > that is taking a long time. Well, I found a reason why ABCL may be taking so extremely long: in SPRDET.lisp, there's a top-level special declaration of the variable name X. In ABCL such declarations are retained after file compilation completes (which is allowed by the spec). Hence, after compilation of sprdet.lisp, any reference to a variable with the name X, or creation of a binding thereof, is a special binding. In ABCL, there's a non-trivial difference between evaluation of a special variable versus a local variable. So, renaming the special variable could be a solution, but that would require a change on part of Maxima; I'm thinking about how to revert the side effect at the end of the file compilation. That's a non-trivial change to ABCL. I haven't figured it out yet. Bye, Erik.

2 5

[armedbear-devel] European Common Lisp Meeting
by Arthur Lemmens 07 Aug '09

07 Aug '09

Hi, We just want to remind you that the registration deadline for this year's ECLM (European Common Lisp Meeting) is August 15. You can still register at http://weitz.de/eclm2009 We hope to see you in Hamburg, Edi Weitz & Arthur Lemmens

1 0

[armedbear-devel] [PATCH] Problems with new JavaObject.javaInstance(Class c) and JSS
by Mark Evenson 05 Aug '09

05 Aug '09

Unfortunately Alessio's (and Tobias?) work on [adding richness to inspection of Java objects][1] contains a bug that manifests itself in stopping [JSS][2] from working. The attached patch gets JSS to run again, but I am not sure the right way to fix the problem in the trunk, mainly as I don't understand all of our conversion rules, and how they interact with Java's autoboxing. The problem with JSS occurs when JavaObject.javaInstance(Class c) is invoked with a Class from a primitive boolean on an JavaObject that wraps a java.lang.Boolean. The call to java.lang.Class.isAssignableFrom() has an oddly (to me) specific contract that "[i]if [the] Class object represents a primitive type, this method returns true if the specified Class parameter is exactly this Class object; otherwise it returns false." This makes JavaObject<java.lang.Boolean>.javaInstance(boolean) fail, when everything seems to work fine if we simply return the wrapped object. What I don't understand mainly is if boolean/java.lang.Boolean objects are somehow special in our treatment of them, or should we do this for all of the other primitive types? [1]: http://trac.common-lisp.net/armedbear/changeset/12081 [2]: http://mumble.net:8080/svn/lsw/trunk/jss/ -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."

3 5