The purpose of the change is to start using ODR hash for comparison
instead of ASTStructuralEquivalence and ODR hash calculation is,
unfortunately, incomplete. The decision to use ODR hash is made because
we can use the same mechanism while checking duplicates encountered
during deserialization. For that case ASTStructuralEquivalence is not a
good fit because it can lead to excessive deserialization and
deserialized decls would be used only for ASTStructuralEquivalence.
rdar://82908223
I'm not sure I'm a fan of using the exact same function name for checking ODR hashes and structural equivalence.