On Wednesday 01 September 2010 11:26:25 am Benson Margulies wrote: > I know I've made some noise about this before, but I'm hoping for some > further clarification before I hare off after a specific sort-of > cosmetic annoyance. > > Here's what recently happened to me.... > > I used the JAX-WS client API to create a proxy. I didn't pass in a > WSDL URL, and there was no WSDL location annotation on the SEI that I > passed in the Class<?> reference for it. > > I confess that I expected it to concoct the obvious ?wsdl URL for > itself, and did not realize that it was deciding to live without. > > What I got was a proxy that worked ... but, on the way, I got a long > list of 'severe' errors from the RFSB. For every complex type used as > a parameter or return from the SEI, out came a severe message > complaining that the relevant complex type was missing.
Well, this is really a problem. It SHOULD have been able to produce a correct schema internally and not output those severe warnings. If it's not producing a correct schema, that really is a problem/bug that needs to be fixed. For example, if we use the same SEI to generate a java first server impl, it would probably produce the same severe warnings and produce an invalid wsdl, which is a problem. Any chance you can log a bug or track down the invalid schema it is creating? Dan > > Presumably, what happened here is that CXF didn't fabricate enough > schema. Yet calls to the service work right. > > It seems to me that these severe log messages can't possibly be what > we want. We want a single WARN level message that says, 'no wsdl, you > idiots, you may have problems.' > > I'd like to work on the code here, and also on the text on Confluence. -- Daniel Kulp [email protected] http://dankulp.com/blog
