I now consider this done, and I believe it can be closed, however, it would still be worthwhile if others would review it.

This is now mostly done (though there’s still some details to finish off).

This was actually way harder than anticipated, and the PRs were rather big, but the good news is I cleaned up a bunch of cruft so that other new things should be much easier to do in the future.

I was thinking an annotation. (By analogy with the @Struct annotation we already have.)

The names can be inferred, and that’s reasonable for exported schemas. But there is still the case of existing schemas.

An implicit naming strategy would be nice at the minimum. Possibly a new “SQL type name” annotation for any explicit cases. Maybe allow to leverage @Column#columnDefinition for the explicit case?

Turns out that supporting enum types on Postgres is way harder than I expected due to how we implemented enum types in Hibernate.

(And also because Postgres implicit type conversions don’t work as well as you would expect.)