Hi, I've been trying to improve my idea in [1]. Attached is a slightly different patch, still with the same general idea though. I've added two tests, however I'd still like to do a bit more than that, especially testing the correct resolution of methods, i.e. Java::isMoreSpecialized.
My current understanding is that with this patch all possible methods (like for the attached class DummyMethods) are found, but that isMoreSpecialized doesn't actually do what its name implies, namely finding the most specialized method.
That is, I'm assuming that JCALL is supposed to use the (exact) same kind of method resolution as Java (if no method signature is given)?
Additionally I actually wanted to add DummyMethods for testing from java-tests.lisp, but didn't work out for now.
Best regards, Olof
[1]: http://abcl.org/trac/ticket/352
_______________________________________________ Armedbear-devel mailing list Armedbear-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/armedbear-devel
On 15 Apr 2014, at 22:57, olof@macrolet.net wrote:
Olof
Applied as [r14682][] and backported to ‘branches/1.3.1’. Thanks for the patch!
I’ll get to the Java test after releasing abcl-1.3.1. If you open things in Netbeans, you should be able to cut ’n paste this code into a test class which we can then reference Lisp-side.
[r14682]: http://abcl.org/trac/changeset/14682
On Thu, Apr 17, 2014 at 01:54:02PM +0200, Mark Evenson wrote:
Applied as [r14682][] and backported to ‘branches/1.3.1’. Thanks for the patch!
I’ll get to the Java test after releasing abcl-1.3.1. If you open things in Netbeans, you should be able to cut ’n paste this code into a test class which we can then reference Lisp-side.
Oh well, Cyrus noticed that this broke loading JNA. Attached is another patch to fix that, including a test case for null vs. primitive in a method signature.
armedbear-devel@common-lisp.net
-
Mark Evenson -
olof＠macrolet.net