il Fri, 27 Jun 2003 06:05:28 +0900, Mauricio Fern?ndez
<batsman.geo / yahoo.com> ha scritto::




>It does now. Static methods seem to become singleton methods of the
>class objects. You can do
>  aklass.java_instance_methods
>to get an array of Method objects, and
>  aklass.java_methods
>to get the static methods (singleton methods) of the class.
<snip of cool samples>
well, I suppose having the java_(reflecting) method separated from the
ruby one could be useful.. but I wonder if someone would prefer full
consistency with "the rest of the world".

Possibly a JRNI::FULL_MIMIC flag for  Reflect.new? 
^_^


>> what kind of name collision could happen ?
>
>classes in Java with methods like object_id, instance_methods, class,
>method_missing, etc...

I suppose that this kind of collision should be handled by the user.
I strongly prefer using 
my_java_obj.method 
over 
my_java_obj.send( :method)

I suppose the name collision for common methods (like 'class') would
be few 
(looking at Class or Object, for example, I can't see a cooomn method)
Moreover, common methods such as Object.class or Array.length would
collide but would give an equal result.. 

BTW, if someone is taking the approach of using a remote Java object
he should be aware if some needed ruby method gets messed up from
java.

Possibly a best way could be aliasing colliding methods such that a
colliding :meth  would make java_obj respond to :ruby_meth and :meth
(or :java_meth and :meth).
Ugly, but this should be really rare imVho

>> anyway I have no idea about the possibility to do this stuff, so
>> forget my desires if you wish :) 
>
>everything doable, by experience: I have done most of it by now :)

Yeah, 
but now I'm waiting for the CORBA binding in the next few days ;)