This is consistently giving me test failures. They are fixed after applying [4/5], so I suspect it's a patch ordering issue.
That's odd, I'm not able to reproduce that at all. There's nothing in the third patch that should cause any difference to the output, and the changes are quite straightforward. It looks like the Publish method is returning E_INVALIDARG in your case. Could you perhaps re-run the discovery tests with WINEDEBUG=+wsdapi set and send me the log?
Thanks,
Owen