> In this case I think the suitable thing to do would be to coordinate work > around > your existing patch via some sort of OSS project and look at all of the > aspects > that would need to be covered in order for it to be accepted into the > OpenJDK. > For example, a lot of thought will need to go into the impact on with > existing and > future concurrency capabilities and it would be good to plan lots of > empirical > testing against a large corpus of existing code.
Some thoughts - I think if we can round up some committed volunteers from the active JVM language communities (e.g. JRuby, Groovy, Scala, Clojure, Jython, Kawa, and the newer ones just arriving), the feedback will be more varied and, I think, more helpful to whatever expert group forms later. One step towards that would be having a downloadable, stable VM with the coro patches available for people to play with. A shared DVCS could be used for people to share their experiments. The shared repository would help, in turn, to verify that ongoing work on Da Vinci doesn't break existing code without people noticing. Patrick _______________________________________________ mlvm-dev mailing list mlvm-dev@openjdk.java.net http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev