Thanks, Laird. As I mentioned we've been discussing this in the EJB
3.1 EG for a couple weeks now, but your note to jsr-318-
[email protected] is still very appreciated and does help show support
for clarifying this in the 3.1 spec.
-David
On Apr 17, 2009, at 8:14 AM, Laird Nelson wrote:
I thought this would be of interest to the list.
Best,
Laird
---------- Forwarded message ----------
From: Kenneth Saks <[email protected]>
Date: Fri, Apr 17, 2009 at 10:45 AM
Subject: Re: JSR 318 Comments
On Apr 15, 2009, at 10:31 AM, Laird Nelson wrote:
Hello, Ken and others.
There's a bit of a thread going on at the OpenEJB mailing lists
concerning
whether or not it is--or should be--legal to use a single business
interface
as both a @Remote interface and a @WebService SEI. Apparently the
specification does not directly address this issue. Do you folks
have a
clarifying statement you'd be willing to share? Ideally one that
would be
retroactive to EJB 3.0? ;-)
Personally what I want to do (portably, reliably) is this (typed off
the
cuff):
@Remote
@WebService(targetNamespace="...")
public interface Calculator {
public int add(final int one, final int two);
}
@Stateless
@WebService(endpointInterface="x.y.Calculator", targetNamespace="...")
public class CalculatorBean implements Calculator {
@Override
public final int add(final int one, final int two) {
return one + two;
}
}
At the moment, if I recall correctly, Glassfish will allow this, I
think,
and OpenEJB will not. I haven't tried JBoss, but, of course, the
whole
point is I'm not sure *who's* correct. I would really, really like
it if
this construct were deemed legal and acceptable.
Hi Laird,
There's nothing in the spec that prohibits this combination of
client views
but it's not a common usage so it's possible there is some
difference in
behavior among vendors. We'll look to clarify the requirements
within EJB
3.1. Thanks for bringing it to our attention.
Regards,
--ken
