Hi!
Dan OConnor wrote:
> > Well, tell that to the EJB spec writers. We follow the specs, which say
> > that an exception must be thrown on reentrant calls to stateful
> > sessions.
> >
> [snip...]
> >
> >
> > See above, the stateful session behaviour may not be changed. At least
> > not in JBoss core. If you want to do local changes, that's fine, but you
> > may not add them to JBoss CVS. Sorry Ole :-(
> >
>
> Hi Rickard & Ole,
>
> In the proposed final draft of the EJB 2.0 specification, page 66,
> footnote (4), regarding multiple concurrent requests to a stateful
> sesion bean, the following is stated:
>
> "In certain special circumstances (e.g. to handle clustered web
> container architectures), the container may instead queue or
> serialize such concurrent requests. Clients, however, cannot rely
> on this behavior."
Excellent! I had missed that one :-) No problemo then.
> Rickard is correct that, from a portable client's point of view, you
> must assume that a RemoteException will be thrown. However, I
> do not think that Ole's proposed change would be a violation of the
> specification. If we do provide queueing behavior for stateful session
> beans, we might want to make it optional. That way, a developer
> who wanted strict portability would not be misled.
Indeed.
> I'm not a big fan of stateful session beans in general. However,
> IMHO Ole's proposed change increases their potential usefulness
> in a web architecture, especially with a stateless load-balancing
> scheme.
Agree, both re: usefulness of stateful beans and Ole's proposal.
/Rickard
--
Rickard Öberg
Email: [EMAIL PROTECTED]
