Thanks for your quick reply.


> 
> 1)
> You can send a dummy NACK response when the client was not supposed to
> get a response
> 
> 

I prefer do not use this solution to maintain proper separation between
usable answers and error ones.
Actually, clients have a result queue (that they supplied in replyTo) which
is dedicated to real answer and an error queue in which my processor will
push a roughly equivalent of your so-called NACK response.



> 
> 3)
> Who is overriding the correlation id? Is it the "private broker" or is
> it Camel itself?
> 
> 

In fact, it's the message id of routed message which is overriden. To affect
correlation id, my processor just a use a simple
out.setJMSCorrelationId(in.getJMSMessageId()) statement.
I didn't look deeply in camel internal gears but i assume routing is
achieved by a jms consumer/producer couple, with jmscomponent configuration
telling jms provider to generate id or not as message sent.



> 
> 2)
> Camel 2.0 have better support for InOptionalOut, so if possible try
> with 2.0. There is a 2.0m1 in the central maven repo.
> 
> 

I'll try to play around this solution.
I didn't before since my activemq version was packaged with a 1.X and 2.X
assume a major api change.

-- 
View this message in context: 
http://www.nabble.com/Routing-keeping-originalMessageId-tp22678749p22680104.html
Sent from the Camel - Users (activemq) mailing list archive at Nabble.com.

Reply via email to