Attempt to use receiver as parent class when it's a call node

[Morriar]

Summary

• When class Foo < Bar is indexed and Bar is never defined, Foo's ancestor chain was silently falling back to [Foo, Object, BasicObject] (partial) — Bar was completely absent. This is inconsistent with how unresolved mixins are handled, which push Ancestor::Partial(name_id) as a placeholder.
• Fixed by applying the same approach for superclasses: when no resolved parent is found but an unresolved reference exists, linearize_parent_class now returns Ancestors::Partial([Ancestor::Partial(name_id)]) instead of falling back to Object's chain.
• Updated assert_ancestors_eq to compare by name string, so it works uniformly for both Complete and Partial ancestor entries (previously panicked on Ancestors::Partial).
• Added Class#superclass_name Ruby API that returns the unresolved superclass name as written in source, or nil when resolved or absent.

Test plan

• New Rust test: unresolved_superclass_is_represented_as_partial_ancestor verifies class Foo < Bar yields ["Foo", "Bar"]
• New Ruby tests: test_superclass_name_returns_nil_when_superclass_is_resolved and test_superclass_name_returns_name_when_superclass_is_unresolved
• cargo test — all 380 tests pass
• bundle exec ruby -Itest test/declaration_test.rb — all 23 tests pass
• cargo clippy --all-targets -- -D warnings — clean

[vinistock]

I think the resolution changes can be shipped, but I do think the superclass API will be nicer returning constant reference objects