Follow up for #38420 We should parametrize over all parsers for the test sets to ensure a consistent behavior