On Thu, Dec 11, 2014 at 08:04:18PM -0500, John Cowan wrote: > Paul Hoffman scripsit: > > > > JSON text sequences MUST use UTF-8 encoding; other encodings of JSON > > > (i.e., UTF-16 and UTF-32) MUST NOT be used. > > > > That seems like a good clarifying addition as well. > > I continue to think this is a bad idea. In particular, you cannot claim What is? > that because a text sequence entity body is not valid UTF-8, that it is > not a valid JSON text sequence; as I explained before, the JSON texts > may be truncated or otherwise damaged such that they contain only part > of a UTF-8-encoded character. RS cannot accidentally result from truncating a multi-byte UTF-8 encoded codepoint. Nico --