yes, there's a tck failure.
-- dims
On Mon, 21 Jun 2004 08:42:56 -0400, Glen Daniels <glen@thoughtcraft.com> wrote:
>
>
> Hi Ias:
>
> > I simply added
> >
> > if (msgContext.getOperation() != null) {
> > // We can test this in a WSDL-centric manner, too.
> > if
> > (msgContext.getOperation().getReturnClass().equals(void.class)) {
> > responseMsg = null;
> > }
> > }
> >
> > in front of sendResponse of /* Send response back along the
> > wire... */ if clause in AxisServlet.doPost() and confirmed
> > that it worked fine by all- tests as well. Please review this change.
>
> -1, if you ask me.
>
> The problem here is that there may be valid request-response operations
> which return void, but might return headers along with the void response.
> If we're going to put in code to enforce this requirement, I think we need
> to make sure that the operation really is one-way. Is some TCK test failing
> due to this?
>
> --Glen
>
--
Davanum Srinivas - http://webservices.apache.org/~dims/
|