When cacluating the content length for XMLRPC replies, the current implementation of XMLRPC servelet just serializes the response into a string and gets its length. This gives the number of characters in the string and not the number of encoded bytes. There is a difference between the two when dealing with characters with multi byte encodings (e.g. kanji characters when encoded in UTF-8), and this results in truncated responses from the XmlRpcServelet when sending responses with such characters.
Patch for fix attached
submitted by nadir@amazon.com (sorry, forgot to log in before submitting this comment)
before submitting this bug, not comment