All, Apologies for the delay. I've been considering what I think about next priorities (as well as having a busy time on return).
I too think an M2.2 would be best avoided unless we encounter any major issues. Here's my conclusions about what it looks like, and I agree with Rob about quite a bit of the work on the less feature related items: Java Broker/Client: Land the 'bigmergebranch' between M2.1 and truck Refactor the broker, as discussed at the F2F on the CSDM side particularly Flow To Disk - to make us able to scale and *not* get OoM Message Priority - proper implementation, requires broker refactor first iirc Message Federation (Fan out & forward to remote queue) Flow control - throttling of overactive producers Exception Handling - argh, needs some work esp on client Management Console Improvements - get it building better & test moving of messages from queues Transparent 0.8/0.9/.10 support I don't think this is miles from anyone else, aside from message priority. I don't have a view about the C++ stuff, aside from that (gap-y as it is) the Java Documentation is more comprehensive that the other implementations and it'd be good to see them catch up a little. Hth, Regards, Marnie
