[1.x] binding.jms correlation ID for callback messages?

[Scott Kurz <sc...@gmail.com>]

Was wondering if we missed something (or consciously decided not to
implement it?)

Do we agree that we're NOT setting the correlation ID per the
correlation scheme for a callback message (i.e setting the correlation
ID on the callback-direction request message based on the
forward-direction request message, depending on scheme)?

I tried a quick test and this seemed to be the case... but wanted to
make sure I wasn't missing something.

When I read the OSOA JMS Binding spec, I see this which suggests we
should be doing this.

114 • /binding.jms/@correlationScheme – identifies the correlation
scheme used when sending
115 reply or callback messages. Valid values are
“RequestMsgIDToCorrelID” (the default),
116 “RequestCorrelIDToCorrelID”, and “None”.

Any thoughts?

Thanks,
Scott

Re: [1.x] binding.jms correlation ID for callback messages?

[Scott Kurz <sc...@gmail.com>]

Here is the OASIS discussion JIRA from my comment:

http://www.osoa.org/jira/browse/BINDINGS-97

Re: [1.x] binding.jms correlation ID for callback messages?

[Simon Laws <si...@googlemail.com>]

OK, fair enough Scott.

Simon

Re: [1.x] binding.jms correlation ID for callback messages?

[Scott Kurz <sc...@gmail.com>]

Thanks Simon,

After considering some more, I'm going to take this up as a spec issue
(rather than argue for us to change our code)... as this sort of thing
I proposed would be conversational-like, and I don't think that's
where we want to go with this.

Scott

Re: [1.x] binding.jms correlation ID for callback messages?

[simonslaws <si...@googlemail.com>]

Hi Scott

Looking at the code...

I don't think we are setting it in the callback case based on what
arrives in the forward message.

I think we are setting it in the reply to case. This is based on how
you configure the JMS binding at the service. It seems that we default
to message ID which may not be right.

I don't think there is a good reason. It looks like it just got missed
out. Want to raise a JIRA? If this is required relatively quickly I'd
be tempted to open a 1.5.2 category.

Simon