Due to the fragmentation of TCP transmission, you cannot reliably expect the 'buf' buffer to always end on a full UTF-8 encoded character boundary. Calling UTF8.GetString(buf) when buf has end-bytes that do not fully make up a proper character can create garbage in the text. Using a UTF-8 Decoder is the proper way to accumulate UTF-8 text over a series of byte buffers. I will submit a pull request for the issue.
Due to the fragmentation of TCP transmission, you cannot reliably expect the 'buf' buffer to always end on a full UTF-8 encoded character boundary. Calling UTF8.GetString(buf) when buf has end-bytes that do not fully make up a proper character can create garbage in the text. Using a UTF-8 Decoder is the proper way to accumulate UTF-8 text over a series of byte buffers. I will submit a pull request for the issue.