I've been exploring the problem and I think we should challenge it.
The way the access type (field, getter) is guessed and thus the annotations processed usually depend on the position of @Id in the class hierarchy. In a MappedSuperclass hierarchy without @Entity, there is no guarantee that we will have a properly defined access type.
So I'd recommend we go and challenge the idea of representing MappedSuperclass in the metamodel unless they belong in a proper entity hierarchy.
Thoughts Steve Ebersole?
|