removed unnecessary Buffer.toString() with hardcoded utf8 encoding

[nlang]

If a message consists of parts with charsets other than UTF-8 (e.g. ISO-8859), converting the buffer to a string with fixed encoding breaks special characters like umlauts. As net.Socket.write() excepts a Buffer as the first parameter, converting it to string is unnecessary anyways, so removing the Buffer.toString() easily fixes that problem.