You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by Marnie McCormack <ma...@googlemail.com> on 2009/09/08 17:53:57 UTC

[Java] 0.6/0.7Upcoming Development

Just a quick update on the 0.6 content & 0.6+ work coming on the Java side,
at least the bits I know about from talking with other java devs.

The Management Console work that Robbie has pulled together should go out as
part of the 0.6 release, along with the Logging Status updates items
completed by Martin which introduces a new operational logging framework.
0.7 work planned includes an extension of this logging to include statistic
generation including rate etc. Also have some plans to work on fixes for a
set of defects affecting failover and acls for 0.7.

(BTW Robbie has recently graduated but fortunately will continue working on
Qpid, which is great news on the Java side.)

The NIO changes for encapsulation that Aidan as been working on are going
well and hopefully we'll see more of the component items from the
design/work breakdown on the wiki make it into 0.7.

Rafi/Rajith/Rob - please fill in the blanks ! Thx.

Anyone else planning items for 0.7 on the Java side ?

I'll be hitting the JIRAs with a stick in little while.

Hth,
Marnie

Re: [Java] 0.6/0.7Upcoming Development

Posted by Martin Ritchie <ri...@apache.org>.
Following Rob's extesnive list of additional work, I only have the one
thing to add :

ACLv2 Plugin for Java Broker

I have implemented the version 2 ACLs for the Java broker, currently
I'm writting a more comprehensive test suite.
Currently the ACLs are only on AMQP connections but it would be nice
to mirror the C++ controls on management objects.

Would be nice to get this in for 0.6 but 0.7 is definately doable.

Martin

2009/9/9 Robert Godfrey <ro...@gmail.com>:
> 2009/9/8 Marnie McCormack <ma...@googlemail.com>:
>
> <snip>
>
>> Just a quick update on the 0.6 content & 0.6+ work coming on the Java side,
>> The NIO changes for encapsulation that Aidan as been working on are going
>> well and hopefully we'll see more of the component items from the
>> design/work breakdown on the wiki make it into 0.7.
>>
>> Rafi/Rajith/Rob - please fill in the blanks ! Thx.
>>
>> Anyone else planning items for 0.7 on the Java side ?
>>
>
> OK - things I have ongoing:
>
> 1) 0-10 Java Broker
>
> I have this passing almost all of the Python 0-10 tests now... but I
> still need to work on integrating the transactional models used in the
> 0-8 codepath and the 0-10 codepath.  Also the configuration for having
> different ports for different protocols needs to be created.
>
> This should definitely be ready for 0.7, though we may want to keep
> the 0-8/0-9 protocol as the default option.
>
> 2) 0-8/0-9 tactical producer flow control
>
> I've got a patch which will allow us to set limits on the size of
> queues after which publishers to those queues will be sent
> channel.flow commands to stop them sending more messages.  The patch
> is complete except for writing some unit tests.  Again this should be
> ready for 0.7.
>
> 3) Last Value Queue
>
> Another patch I've had hanging about for a while to allow for a queue
> which retires entries when a new message with the same "key value"
> enters the queue.  It's a bit dusty but shouldn't take long to get
> back into shape.
>
>
> I've also done some tentative work in the same direction that martin
> previously went on Flow-to-Disk... but that's very much more
> experimental.  Similarly I've got an experimental message store
> floating around somewhere which was showing approx 3x the performance
> of BDB... but not likely to be in 0.7.
>
> One thing I would like to examine sometime once we have the Java
> Broker and C++ broker speaking the same version of AMQP is bringing
> the configuration of the two brokers closer... to make everything look
> more like being part of the same product rather than two completely
> independent projects...
>
> Think that's all that I have lying around :-)
>
> Cheers,
> Rob
>
> ---------------------------------------------------------------------
> Apache Qpid - AMQP Messaging Implementation
> Project:      http://qpid.apache.org
> Use/Interact: mailto:dev-subscribe@qpid.apache.org
>
>



-- 
Martin Ritchie

---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project:      http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org


Re: [Java] 0.6/0.7Upcoming Development

Posted by Robert Godfrey <ro...@gmail.com>.
2009/9/8 Marnie McCormack <ma...@googlemail.com>:

<snip>

> Just a quick update on the 0.6 content & 0.6+ work coming on the Java side,
> The NIO changes for encapsulation that Aidan as been working on are going
> well and hopefully we'll see more of the component items from the
> design/work breakdown on the wiki make it into 0.7.
>
> Rafi/Rajith/Rob - please fill in the blanks ! Thx.
>
> Anyone else planning items for 0.7 on the Java side ?
>

OK - things I have ongoing:

1) 0-10 Java Broker

I have this passing almost all of the Python 0-10 tests now... but I
still need to work on integrating the transactional models used in the
0-8 codepath and the 0-10 codepath.  Also the configuration for having
different ports for different protocols needs to be created.

This should definitely be ready for 0.7, though we may want to keep
the 0-8/0-9 protocol as the default option.

2) 0-8/0-9 tactical producer flow control

I've got a patch which will allow us to set limits on the size of
queues after which publishers to those queues will be sent
channel.flow commands to stop them sending more messages.  The patch
is complete except for writing some unit tests.  Again this should be
ready for 0.7.

3) Last Value Queue

Another patch I've had hanging about for a while to allow for a queue
which retires entries when a new message with the same "key value"
enters the queue.  It's a bit dusty but shouldn't take long to get
back into shape.


I've also done some tentative work in the same direction that martin
previously went on Flow-to-Disk... but that's very much more
experimental.  Similarly I've got an experimental message store
floating around somewhere which was showing approx 3x the performance
of BDB... but not likely to be in 0.7.

One thing I would like to examine sometime once we have the Java
Broker and C++ broker speaking the same version of AMQP is bringing
the configuration of the two brokers closer... to make everything look
more like being part of the same product rather than two completely
independent projects...

Think that's all that I have lying around :-)

Cheers,
Rob

---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project:      http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org