Check UniversalType's equals and hashCode methods

[ljnelson]

The hashCode method simply returns the delegate's hashcode. I can't see how this could be wrong.

The equals method currently uses isSameType semantics, which I'm thinking may be incorrect. The problem seems to surface in certain type erasure scenarios where two delegates, from the same implementation, representing to the naked eye java.lang.Object, do not compare as the same type (!) so the equals call fails.

[ljnelson]

This was good to catch. UniversalType#equals(Object) was implemented in terms of sameType, but of course there was no way for hashCode to be similarly implemented, so you had a classic case of violating the hashCode contract. I will be ensuring that equals just goes to the delegate.

This has some follow-on effects: consumers may be surprised that certain TypeMirror instances no longer compare as equal.