Correction and simplification of `TypeDetails#isResolved()` implementation

[mbladel]

After what we found about the expectations for TypeDetails#isResolved() is ORM, I believe that we can simplify our current implementation of it while also ensuring correct semantics:

• Class types should always be considered resolved, even java.lang.Object (we already do this for "first level" types, e.g. Object typed members)
• Wildcard types with class-typed bounds should be considered resolved (we already do this with a special case for TypeDetails.Kind.CLASS)
• Type variables should always be considered not resolved, as the type is always dependent on the specific class' type parameters, regardless of the type variable bounds
• The same goes for type variable references