BytesToUTF8 has incorrect behavior when buffer contains incomplete code point

[frank]

rgl says: http://orbited.org/browser/branches/0.5/daemon/orbited/static/protocols/stomp/stomp.js#L21 does not seem to work when a unicode code point is a multi-byte one and we receive them in multiple calls; I think the translation from bytes to unicode should be made after we detect the end of headers and the end of frames.

[heyadayo]

Solved in 0.6.0 (somewhere between r470-491)

function names are now Orbited.utf8.toUtf8 and Orbited.utf8.fromUtf8