David is much more likely to know the right answer than I am with regard to whether that's legal, but it looks like the webservice doesn't deploy in OpenEJB correctly with an SEI.
Sorry that's not much help - I'll have a play around with this and see if I can get OpenEJB to use one interface for both the webservice and remote interfaces hopefully later on this week. Jon On Tue, Apr 14, 2009 at 8:43 PM, Laird Nelson <[email protected]> wrote: > On Tue, Apr 14, 2009 at 3:31 PM, Jonathan Gallimore < > [email protected]> wrote: > > > I remember thinking this was a pain when I hit it about a year ago - I'm > > glad its not just me :) > > > Same here! > > > > As David says, we could probably make this work, but its against the > > current > > spec, so the code might not work with other containers. I'm happy to have > a > > go at this when I have some free time (unless someone else beats me to > it) > > ;-) > > > Right. > > Now, it is within the specification boundaries to have a web service > implementation class not reference any SEIs (i.e. you just put @WebService > on the class, and that's that). Would it be legal to have such a class > implement, say, a @Remote business interface only, but also be > annotated--just itself, no SEI--with @WebService? > > It strikes me then that clients could generate their own stubs without > having the hassle--is this the only place in the entire Java ecosystem > where > the presence of an interface is a hassle? Could be!--of dealing with the > SEI. > > Best, > Laird >
