reject overlong UTF-8 sequences for which a shorter encoding exists.          |