On 28/09/2010, at 3:12 AM, Dan Allen wrote:
Stuart has mentioned in several threads that fakereplace is very close to JRebel in core features, and in some cases perhaps farther along. The big thing that JRebel has is the integrated tooling support and a product w/ support. I applaud their work, but am still very much in support of Stuart's open source and extensible alternative.
I would not go that far :-) The guys at JRebel have spent years working full time on it, Fakereplace has not made it much past the prototype phase.
Architecture wise I think what is really needed is a JSR that will allow frameworks and hot deployment providers to work together in a loosely coupled way. That way frameworks can write the code to perform any re-initialistion of metadata that is required, and this will work with any hot deployment provider. If the JVM replacement functionality advances far enough Fakereplace's bytecode tricks will no longer be necessary.
Stuart
One way or another, we should definitely be in this space. Perhaps if Stuart finds what the JVM guys are working on interesting, he will be willing to participate or provide feedback. From my understanding, you still need framework support and fakereplace will provide that architecture. So, I'd say that language level would be great, but still just one factor in the whole equation.
_______________________________________________
seam-dev mailing list
seam-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/seam-dev