Intercept the JMSMessage before sending it? pairing up clients and service instances.

[miha]

Hi all,

I'm trying to figure out if the scenario in the title is even possible?
The reason i'm asking is the following:
1. We have multiple service clients - all SpringWS implementation
2. We have multiple service instances running on the same request/response QUEUES, but on different machines.
3. We want only some clients to be able to invoke some service instances....

My idea is to use JMS messageSelectors for filtering, which is possible on the DMLC (server side) but i'm lacking a mechanism which would enable us to set the custom header property on a JMS message on the client.

Any best practices with these kind of scenarios?

Thanks in advance!

[Arjen Poutsma]

You can get a reference to the request message by using the TransportContext. See http://static.springframework.org/sp...nsport-context, both on client (in a callback) and on server.

Cast the connection to a JmsSenderConnection, and get the messages from there.

Let me know if that works for you

[miha]

...I should have read the docs more carefully - i wouldn't have missed the part you pointed out....
Off course it works, like a charm. This is the missing piece in our multi-instance clients and services development setup...

Thanks!

[miha]

...seems that the Message retrieved via the TransportContext returns back as NULL...

Which leads to NPE when trying to mess around with the Message....I'll try to investigate a bit...
Sorry for the hassle...

[miha]

...that the

Callback.doWithMessage

executes before

JmsSenderConnection.onSendBeforeWrite

which is the method that actually creates the

javax.jms.Message

Could the creation of the Message be dealt with earlier in the execution chain?

Thanks in advance!

[Arjen Poutsma]

Could the creation of the Message be dealt with earlier in the execution chain?

Yes, I have created http://jira.springframework.org/browse/SWS-311 to fix this.

[miha]

I've playing around with the JmsSenderConnection class, would you mind if i attached my temp solution to the JIRA issue you created?

[Arjen Poutsma]

It's fixed. If my solution is different than what you had in mind, I'm all ears .

[miha]

...until it becomes available in a 1.5 snapshot or in the final release, we'll be using our own

Thank you very much for such a quick reaction!

P.S. Our patch is just a bit different, i would appreciate if you took the time to look at our solution (in the attached archive) and comment it..And point out the errors of our way