that we also check to confirm both engines have consume the same number
of tokens (and thus have actually both arrived at the same state for the
same partial input stream)