Hi Aki,
Yes, you understand my scenario and my struggles well. The setup you
describe does seem to be an option.
client(wsdl1) --> EP1(wsdl2)/trans--camel-route- EP2(wsdl2)-->
service(wsdl2) 
The only issue i have with this is that, while the incoming request (WSDL1
versus WSDL2) is pretty similar, the response from backend service (WSDL2)
is quite different from the response expected by client (WSDL1). 
Response requires quite a bit of manipulation. After invoking the backend
service, the response adhering to WSDL2 propogates back through my route.
Transforming the response from WSDL2 to WSDL1 requires some enrichment from
external system and more work than the CXF Transform feature is capable of.
So much manipulation that I'd like to be able to perform some of this from
the camel route - after the call to backend service (but before message
passes back through EP1 and response sent back to client.) 

With the setup you describe, If I do this tranformation in Camel route, I
assumed things would blow up when modified response hits EP1 on the way back
to client, but i found that is not the case. The EP1 will pass the response
back (malformed according to WSDL2) to client  without the operation/binding
mis-matching I see on the way in if i send a malformed request. So i think
the solution you suggest works for my scenario. Thanks!

I'm curious, In order to use of CXF transform easier within Camel.. Is there
any way to disable requirement that operation/binding match in CXFEndpoint
on incoming requests? That way i could disable this requirement at the
Endpoint so i could freely apply tranform into any format at the endpoint
and it wouldn't blow up. Otherwise I can only transform into something that
conforms with the WSDL the endpoint is tied to. Although in doing so I
realize I may be abusing somewhat the use-case this feature is intended for.
 Thanks,
-Lowry




--
View this message in context: 
http://cxf.547215.n5.nabble.com/cxf-transform-feature-for-Declarative-transform-in-camel-route-tp5746295p5746339.html
Sent from the cxf-user mailing list archive at Nabble.com.

Reply via email to