Zarar,
Just back from my trips and looking at things....
I'm not sure if the patch is fixing the correct issue. According to the
WSDL spec (section 3.6), the SOAP faults are always "document" style,
thus the part needs to be defined as an "element", not a "type". Thus,
the namespace should ALWAYS be properly found from the namespace of the
schema that defines the element.
The code:
if (mpi.isElement()) {
ns = mpi.getElementQName().getNamespaceURI();
} else {
ns = mpi.getTypeQName().getNamespaceURI();
}
The "else" clause should never be reached.
If CXF is generating WSDL's that have fault parts defined with "type",
that's a different bug and would need to be fixed. Does that make
sense? What is probably happening is the aegis databinding or one of
the frontends or something is likely creating faults as types instead of
elements. Not really sure though. I'd need to see a simple test
case. (Your patch didn't have a test that shows the complete issue)
Dan
On Friday 14 September 2007, [EMAIL PROTECTED] wrote:
> Hi,
>
> I can patch this if you just tell me whether my thinking makes any
> sense.
>
> https://issues.apache.org/jira/browse/CXF-1003
>
> Thanks,
> Zarar
--
J. Daniel Kulp
Principal Engineer
IONA
P: 781-902-8727 C: 508-380-7194
[EMAIL PROTECTED]
http://www.dankulp.com/blog
