Hi Mark.

I have to tell you that the anonymous class loader is a purely experimental 
feature, with no standard API.  For this reason, you can only use it reliably 
with MLVM builds.

To run on a standard JVM, your bytecodes need to be loaded under a name.  It 
can be an arbitrary name, of course.

For purposes of further experimentation, I suggest doing a portable 
implementation that doesn't rely on anonymous classes, and switching to it if 
you get a linkage error trying to use the anonymous class feature.

What advantage does the use of anonymous classes gain you?  Is there a 
performance advantage?

The http://asm.ow2.org framework is an excellent way to spin bytecodes.  This 
framework has an "asmifier" tool which takes pre-existing template-like classes 
and sets them up for modification, in a way similar to the CP patching feature 
of anonymous classes.

Should we consider anonymous classes  for a future version of JSR 292 (or 
another JSR)?  I've been looking for experience to justify standardizing them, 
but I don't see much yet.  Method handles give more flexibility and similar 
optimizability.

-- John


On Jul 11, 2011, at 5:18 PM, Mark Roos wrote:

> Hi all 
> 
> I have be using sun.invoke.anon.AnonymousClassLoader for generating my 
> classes with 
> no issues until now.  With the 6/30 MacOs version from google code I find 
> that my code runs fine but when 
> I go to recompile it I get the package not found error.   I see the package 
> in the rt.jar using a file viewer so its 
> there.  No problems with Stephen's versions. 
> 
> Should this be in the distribution by default? Do I need some compiler flag 
> to enable it? Or is this a bug I should report?   
> 
> thanks 
> 
> mark 
> 
> _______________________________________________
> mlvm-dev mailing list
> [email protected]
> http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev

_______________________________________________
mlvm-dev mailing list
[email protected]
http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev

Reply via email to