Re: [public-ws-addressing-tests] <none>

Glen Daniels wrote:
> 1143 and 1243 are failing because MS is sending a non-URI value for the
> duplicate Action header, and Axis is catching that as a problem before
> it notices the invalid addressing header fault.  Since this is
> essentially a schema validation error, I think we're OK in our behavior
> (what do others think?) - that box will turn green if MS can change the
> "duplicate-action" value to something like "http://duplicate-action".

This was also an issue we had with the Axis endpoint, apologies for not 
mentioning it to the list.  Changing to a proper uri fixed the issue.

> It looks like 1142/1242 are failing because we're trying to send the
> fault to http://duplicate/ (the value of the second FaultTo).  What's
> the right behavior here?  Should we always default to anonymous for
> duplicate FaultTo/ReplyTo errors?  It certainly doesn't say that in the
> spec anywhere.... is this something we need to clear up in the spec?

Our stack also assumes that any fault raised during the processing of 
the WS-Addressing headers, however it is caused, will result in a fault 
being sent back in the response (i.e. anonymous).

It would be nice if the spec did clear this up.

> So if MS fixes the three client issues, that'll clear 1143/1243/1170,
> I'm looking for guidance on 1142/1242, and hopefully we can fix
> 1150/1250/1251.

This matrix looks better every day :-)

	Kev

Received on Thursday, 9 March 2006 10:22:34 UTC