Should we enforce a ServiceRegistry-to-SessionFactory model?
Or, should a single ServiceRegistry be usable in multiple SessionFactorys?
On Wednesday, February 09, 2011, at 09:30 am, Steve Ebersole wrote:
I agree.
On Wednesday, February 09, 2011, at 01:21 am, Max Rydahl Andersen wrote:
> On Feb 9, 2011, at 07:55, Adam Warski wrote:
> > Do you have in mind the process of discovery (that services would also
> > be discovered via META-INF/services, like event listeners), or that
> > the event listener registry should just be another Hibernate service?
>
> If that happens I hope we will have a way to explicitly control it per
> sessionfactory since not every sessionfactory would like every service
> enable as available from the global classpath.
>
> i.e. today from tools we explicitly disable search and validator to avoid
> classloading conflicts and other funky sideeffects when running code
> generation or prototype queries.
>
> /max
>
> > Adam
> >
> > On Feb 8, 2011, at 6:49 PM, Steve Ebersole wrote:
> >>
http://opensource.atlassian.com/projects/hibernate/browse/HHH-5562
> >> discusses applying the principle of a "service locator pattern"
to
> >> event listener registration from integrators. There has also been
> >> discussions about applying the same principles to services as they
> >> will be defined in Hibernate 4[1]. Personally I am leaning towards
> >> combining these into a single mechanism where the LocatedThing could
> >> do either (or both). Looking to see if anyone has specific use cases
> >> where that will not work, requiring the split.
> >>
> >>
> >> [1]
> >>
https://sourceforge.net/apps/mediawiki/hibernate/index.php?title=Categ
> >> or y:Services
> >>
> >>
> >> ---
> >> Steve Ebersole <steve(a)hibernate.org>
> >>
http://hibernate.org
> >> _______________________________________________
> >> hibernate-dev mailing list
> >> hibernate-dev(a)lists.jboss.org
> >>
https://lists.jboss.org/mailman/listinfo/hibernate-dev
---
Steve Ebersole <steve(a)hibernate.org>
http://hibernate.org