The decodeEntities function in xml-enumerator re-invents the wheel that already exists in the text package for decoding character entities. As usual in such cases, the implementation is not only much longer and more complicated, but also incomplete. For example, an entity whose value is invalid for Unicode will cause the Enum instance of Char to call error instead of raising an XmlException. Attached is a patch that uses existing library functions to do the conversion correctly. Regards, Yitz