@jouborg *proper tail calls, not tail call optimization. The latter is a discretionary improvement, the former is a space usage guarantee.