Fur, can you be my Obi-Wan here? /be
I did take a brief look at this. There's one issue that needs to be addressed in the C-engine that does not apply to Rhino: What classloader should the interface-implementing proxy class use ? (It shouldn't use the system classloader as that would give it privileges.) If we want to limit ourselves to JDK 1.3, there's java.lang.reflect.Proxy, a new class that does 99% of the work for us. It will generate a class for an interface and delegate any interface methods to a proxy object, passing in a Method object and an array of arguments. See http://java.sun.com/products/jdk/1.3/docs/api/java/lang/reflect/Proxy.html
Setting all Javacript bugs to rginda QA Contact.
George, is this Blackwood stuff that you guys would prefer to have working?
Uhhhh, I dunno. We'll look at it after Beta 1.
Adding beard, who may be uniquely qualified to help here. Patrick, any thoughts on getting this into SpiderMonkey? See "John Smith"'s (yah right) posting at news://news.mozilla.org/8bcqjk%243fc3%40secnews.netscape.com -- but he's not the only one who craves this fix. /be
I don't think we'd want to limit ourselves to JDK 1.3, as that would leave a lot of developers out in the cold (e.g. the Mac, and users of IBM's VMs?). Perhaps we should study java.lang.reflect.Proxy, and use it if available. However, our JavaAdapter would be very suitable for implementing this, simply using JSObject for the delegation, rather than calls into Rhino's ScriptRuntime. I'd be happy to own this bug.
I think this should be js1.5 (and therefore, JS1.5 won't release till after you fix it in M18 timeframe). Tell me if you disagree, and in any event, what's the latest thinking on JS1.5 release date? There are still a few bugs open, one of them mine for M16! /be
This will have to wait, or land on our post beta3 branch.
Updating QA contact -
-> default qa
I don't think we're going to fix this. reopen if you disagree.