Sebastien,
I should say that I'm working on a default binding that replaces the
Tuscany implementation. The URI isn't changed but new references are
created with target URIs that are unknown to the binding implementation
on the service side.
CompositeBuilderImpl.createSelfReferences creates new references for
each service and sets the name to $self$.<servicename>. Later
wireComposite iterates over each of the references and finds the new
$self$ references and associates it with the default binding.
On paths where SCARuntime.getService is called, createSelfReference
eventually follows and the Reference passed back to the caller is one
that holds generated uri. As the reference's target URI doesn't point
to a service exposed at that URI, invocations through the reference are
failing.
I can see how I might work around this in the binding but I'm not sure
that I should have to. ;)
Thanks.
Jean-Sebastien Delfino wrote:
Matthew Sykes wrote:
I'm actually bumping up against one of the problems that ant has
described - the creation of $self$ reference URIs.
I don't believe a binding implementation should have to expect to deal
with odd URI's that were generated by the runtime and I'm curious how
bindings are supposed to know what to do with each of forms that are
invented.
In the case of the binding implementation I'm working with, the
reference binding uses the uri to determine the target service. When
the URI is changed it is no longer valid as no service exists with at
the modified URI.
I'm trying to understand the problem that you're running into, and I'm
not sure why the URI would change. A new "self" reference created in
CompositeBuilderImpl.createSelfReferences actually points to the
original bindings of the service that the reference is created for. So
as far as I can tell, the URI used by the reference is the same (meaning
exactly the same object) as the URI declared on the service's binding.
So if the URI changes there's a bigger problem, as it's probably going
to break the service itself... Are you actually seeing the URI change?
do you have a test case that I could use to debug that?
Thanks.
--
Jean-Sebastien
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
--
Matthew Sykes
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
