I can't remember if I brought this up a few months back or not...
but I question whether in DataTransformationInterceptor we really need to go
as far as to match the TypeInfo of the source and target fault DataType
logicals for XMLTypes.
In DataTransformationInterceptor.typesMatch() we do:
return matches(t1.getElementName(), t2.getElementName()) && matches(
t1.getTypeName(), t2.getTypeName());
Wouldn't just matching the elem names be enough?
If we really needed to match the types then we have to worry about WSDL
introspection matching xxxExceptionHandler.getFaultType(). E.g. in the
case I just ran into
the WS binding was using WSDL introspection and this was being compared
against my impl's intf built by Java introspection which ended up calling
SDOExceptionHandler.getFaultType().
We'd even have to worry about handling anonymous types the same way. (I
wonder, do we consistently use 'null' as the DataType TypeInfo for anonymous
types throughout our runtime??)
Any thoughts?
Scott
