I think that is just an oversight back to the original impl of the JPA metamodel. I cannot speak for the tests. I can definitely see the argument for PluralAttribute#isAssociation logically returning false for @ ElementCollection.
The spec is very unclear about it unfortunately. WRT entities actually the term I see more often in the spec is "relationship". Personally I logically segment collections exactly the way you suggest, so +1 from me. The only really consideration I guess is compatibility. But still, +1 from me. On Thu, Jun 23, 2016 at 4:37 PM Gail Badner <gbad...@redhat.com> wrote: > Currently, Hibernate's implementation of > javax.persistence.metamodel.PluralAttribute#isAssociation always returns > true. I haven't found anything in JSR 338 (yet) that refers to an element > collection as an association. > > I don't see a good definition of "association", but the term is used with > respect to entities. I also see examples of @ElementCollection with > @AttributeOverride, but none with @AssociationOverride. > > The change suggested in HHH-10888 is to have > PluralAttributeImpl#isAssociation return true only for one-to-many and > many-to-many. I believe this is correct, but it breaks a couple of tests > that specifically check if an element collection is an association. > > I wanted to pass this by you in case I'm missing something here. > > Please take a look at the pull request to see the fix and the required > test changes. [1] > > Thanks, > Gail > > https://github.com/hibernate/hibernate-orm/pull/1447 > _______________________________________________ hibernate-dev mailing list hibernate-dev@lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev
