Robert Godfrey wrote:
Hi All,
prompted by Aidan's earlier e-mail, I guess I should be giving an update on
where I am at with the implementation of AMQP0-10 in the Java Broker
The current status is that the broker passes all the AMQP0-8/0-9 Java tests,
the AMQP 0-8/0-9 Python tests, the Java tests running against the Broker in
0-10 mode, the 0-10 Python tests (after changing some of the Python tests
where I think they are depending on non AMQP 0-10 compliant behaviour in the
C++ broker.
Further the Java Broker can successfully translate messages between AMQP
0-8/0-9 and AMQP 0-10 (i.e. a message sent from a 0-8 client can be received
as a 0-10 message by a 0-10 client and vice versa).
Overall I think it is looking in pretty good shape.
I'm aware of the following tasks as still being outstanding:
1) Allow 0-8/0-9/0-10 to all listen on the same port (currently I'm running
0-10 on a separate port)
2) Update the config to allow for multiple ports, with different protocols
running on them
3) Ensure that the JMX management features are equally available on 0-10
connections as they are on 0-8/0-9 connections
4) Ensure that the logging work recently introduced by Martin is copied over
to equivalent classes dealing with 0-10 connections
Further I need to merge work done on Aidan's network io refactor branch, and
on trunk from the last couple of weeks.
I'm looking to get everything working on a the same port tonight... the
other tasks are less high priority I think.
This sounds very encouraging. It would be good to have the Java Broker
speaking 0-10 for the upcoming release. Do you think it would be
reasonable to land the branch in time? I imagine it would be preferable
to get it in and stabilized rather than let it diverge over the course
of the release.
--Rafael
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:[email protected]
