Ant,
I think I see what I was misunderstanding now...
To set up the DTI (or not) we'll compare the ref. binding IC with:
reference.getReference().getInterfaceContract()
not merely:
reference.getInterfaceContract()
The former, I believe Raymond has explained, is the "componentType" IC while
the latter is the component IC.
OK, so this implies that we will honor the WSDL intf on an component
ref/service
<interface.wsdl> and do a data transform using the WSDL intf format. If
the WSDL is doc-lit-wrapped we'll
use the wrapper elem in the transform.
That last bit is what I was really wondering...
I'm going to try to set this up before asking any more questions.
Thx, Scott
On Fri, May 9, 2008 at 10:49 AM, ant elder <[EMAIL PROTECTED]> wrote:
> On Wed, May 7, 2008 at 3:59 PM, Scott Kurz <[EMAIL PROTECTED]> wrote:
>
> > I have a question on the JMS binding. If I were set up I'd just
> > experiment,
> > but let me just ask....
> >
> > It looks to me like we're requiring the service client/impl to use
> > OMElement
> > as its programming model (i.e. app databinding).
> >
> > I say this looking at class JMSBindingReferenceBindingProvider and
> method:
> >
> > protected void setXMLDataBinding(RuntimeComponentReference reference)
> {
> > try {
> > InterfaceContract ic =
> > (InterfaceContract)reference.getInterfaceContract().clone();
> >
> > Interface ii = (Interface)ic.getInterface().clone();
> > ii.resetDataBinding("org.apache.axiom.om.OMElement");
> > ic.setInterface(ii);
> > reference.setInterfaceContract(ic);
> > ...
> >
> > This looks to me like it will "turn off" the
> DataTransformationInterceptor.
> >
>
> I don't understand why that would "turn off" the
> DataTransfomationinterceptor? Doesn't this tell it to transform the
> payloads
> into an OMElement? (And that does seem to be what happens when i debug
> through the JMS itests)
>
> ...ant
>
