Rajika Kumarasiri wrote:
On Mon, 2008-10-20 at 13:00 -0400, Carl Trieloff wrote:
Rajika Kumarasiri wrote:
On Mon, 2008-10-20 at 11:52 +0100, Robert Godfrey wrote:
2008/10/20 Aidan Skinner <[EMAIL PROTECTED]>:
Has anybody started on that? I know there's been some work done on
making it easier to do with the broker. Are there any design docs?
My refactoring of the broker took into account some of the things we
will encounter in 0-10, such as changes to flow control and a distinct
"acquired" state for message entries.
I think the biggest hurdle will be to re-engineer the broker to handle
the 0-8 and 0-10 stuff without making it such a dogs breakfast as the
client. I really don't want to see the same level of separateness in
the broker if at all possible.
I think I would start by getting the 0-10 style I/O stuff into the
broker and removing the remaining MINA-isms ... We can throw in the
0-10 stack but then need to get some sort of model of Channel/Session
that fits comfortably in the broker and works with the 0-8/9 and 0-10
dispatch mechanisms.
So the issues that are reported against Java broker need to fix in this
process, right ?
-Rajika
Best is to hit Rob and Rafi up if you want to help out with the 0-10
broker work. They have spent
the most time looking into how best to do it as far as I know. I am sure
they will be glad for any
extra help they can get.
Yes, I would like to contribute to the Qpid project and get Qpid java
broker 0-10 specification support. Any guide on the JIRAs which should
be get fixed for 0-10 specification support of Java broker would be
highly appreciated. Thanks!
I can't point you to any specific JIRAs at the moment. Right now I'm
working on getting my coding done for the M4 freeze, but as soon as
that's over I think we should put our heads together and work out a plan
for getting the java broker to 0-10. Hopefully then we'd be able to
point you in a more useful direction.
--Rafael
