Date: Wed, 20 May 2009 09:08:02 -0700
From: Scott Lewis <
slewis@xxxxxxxxxxxxxxxxx>
Subject: Re: [ecf-dev] Re: pub-sub ActiveMQ remote services through
RFC 119
To: "Eclipse Communication Framework (ECF) developer mailing list."
<
ecf-dev@xxxxxxxxxxx>
Message-ID: <
4A142AE2.4070904@xxxxxxxxxxxxxxxxx>
Content-Type: text/plain; charset=UTF-8; format=flowed
Hi Roshan,
roshan joseph wrote:
> Hi Scott,
> Thank you for the detail response and it was very
helpful. I still
> have some questions on the test case.
> I was thinking that, to test the activemq pub-sub remote services
> through RFC 119 you need to have 4 processes, they being
> 1. The activemq broker which creates the topic for pub-sub (for e.g:
> tcp://localhost:61616/exampleTopic)
>
> 2. An ActiveMQJMSServerContainer (ActiveMQ Manager) which has the id
> ‘ecf.jms.activemq.tcp.manager’ and manages the pub-sub group.
>
> 3. An ActiveMQJMSClientContainer (ActiveMQ Client) which has the id
> ‘ecf.jms.activemq.tcp.client’ that acts as the publisher of remote
> services
>
> 4. Another ActiveMQJMSClientContainer (ActiveMQ Client) which also has
> the container id like 3 above but who is the consumer of this service.
>
> Is my thinking wrong here, or can this steps 2,3 and 4 executed in a
> single
process?
>
Yes, that's true. I haven't yet been able to create a new test case
with the above scenario. Since you have it clearly in mind, perhaps you
could create such a test case? If you would like/able to do this,
please open a bug describing this scenario and we'll coordinate on the bug.
>
> I was able to execute the test case under
> org.eclipse.ecf.tests.remoteservice.AbstractConcatHostApplication and
> the AbstractConcatConsumerTestCase both the r-osgi and the generic
> container tests. For the pub-sub using activemq provider won't the
> case be a bit different, as here both the publisher and consumer are
> clients itself of Activemq manager.
>
Yes the scenario is a little bit different. I just haven't been able to
get to creating this new test case yet (as you probably have noticed,
working on documentation,
examples, testing, and support via the mailing
list and newsgroups...as well as giving the OSGi/RFC119 folks a hard
time for changing the spec on us...have all been on my mind recently :).
>
> I will be more than happy to contribute from my side any tests with
> the respect to this topic. If you feel I am way off mark with regard
> to how pub-sub mechanism works in ecf, pls do correct me
>
No afaict you are correct. If you can/could help out with test cases
for this exact scenario it would be most appreciated. If not I will get
to it, but it won't/can't be immediately.
Scott
>
> Regards
> Roshan
