I don't see the impact on setNull either, and yet Table B-5 in the spec clearly states that its tabulation of Java Object Types applies to both setObject and setNull. That reference to setNull in the Table B-5 descriptive language confuses me, as setNull only cares about SQL types, not Java Object Types, I think.
The JDBC 4.2 spec seems quite clear, in Table B-3, that getObject will never return one of the new Java types; it returns the existing Date, Time, or Timestamp types only.
However, the variants of ResultSet.getObject which take a Class instance as an argument might potentially be called with, e.g., java.time.LocalDateTime.class, or even with java.time.OffsetTime.class? In the first case, it seems like we ought to be able to implement this behavior, should we desire to, while in the OffsetTime case, it seems like we ought to perhaps have some sort of "not implemented" behavior?