> Looks good. Andr=E1s, let me know what you think you'll be working on
> so we don't duplicate effort.
Given my very limited Java skills, I think I'd better just stick to
beta-testing and trying to fix/fill small bugs/holes. I also plan to
test compatibility with jlinker. So please just go ahead and do what
you think should be done.
> One thing that seems to be missing from the Franz API is a way to
> query a class and get field lists, method lists, etc. I was thinking
> of something like:
> (jclass-fields) =3D> array of field descriptors
> (jclass-methods) =3D> array of method descriptors
> A "field descriptor" would be like '("field1" "boolean").
> A "method descriptor" would be like '("myMethod" "java.lang.String" "int"=
> Sound workable? I'll try to hack on it this weekend if I have the time.
Reading this, at first I thought, well, there's always the javadoc to
turn to, but now I suspect that what you propose would allow for some
cool dynamic hacks, like automagically defining clos classes/methods
for java classes. That would be great. Ideally, the lisp user
shouldn't ever write low level stuff like jcall & friends. Is this
what you have in mind? Is this possible at all?
Speaking of introspection (and this has nothing to do with the Java
bridge), I think that providing enough lisp functionality to satisfy
ilisp's or slime's needs would make lisp hacking in AB a lot more
enjoyable. All right, you probably all use J :-), but a decent lisp
mode in J probably needs more or less the same stuff.