[jboss-user] [JBoss AOP Development] New message: "Re: ClassPool Refactoring"

Kabir Khan do-not-reply at jboss.com
Fri Jan 8 06:18:11 EST 2010


User development,

A new message was posted in the thread "ClassPool Refactoring":

http://community.jboss.org/message/518925#518925

Author  : Kabir Khan
Profile : http://community.jboss.org/people/kabir.khan@jboss.com

Message:
--------------------------------------------------------------
> alesj wrote:
>  
> > Finally, I detected a great deal of calls to ClassPool.get("org.jboss.aop.microcontainer.annotations.DisableAOP"). The calls are being made by org.jboss.annotation.factory.javassist.DefaultValueAnnotationValidator.getDeclaredMethods(Class<?>), and I'm confident that this can be improved somehow (such as caching DisableAOP ctClass). I think improving this will have some impact on the startup time of AS.
> We're deprecating this annotation anyway.
> But we're replacing it with just the opposite - EnableAOP, so not a lot different from performance perspective.
> Kabir, any idea on how to optimize (e.g. cache) this?
 
The quick and dirty way would be at DefaulValueAnnotationValidator level
-a WeakHashMap<Class<?>, CtClass>, or
-be able to configure it with annotations it should cache
 
But I guess the real and harder solution is that caching at class pool level could be improved? What seems to happen now is that when trying to load a CtClass we end up in Base-/JBossClClassPoolDomain which checks each relavant pool and parent domain for a locally cached CtClass which is simple but probably more work than necessary. Maybe these lookups should be cached in the initiating pool, with some invalidation if the domain (or its parent domain) has pools added/removed?

--------------------------------------------------------------

To reply to this message visit the message page: http://community.jboss.org/message/518925#518925




More information about the jboss-user mailing list