scan_utf8 decodes an unsigned integer in UTF-8 encoding from a memory
area holding binary data. It writes the decode value in dest and
returns the number of bytes it read from src.
scan_utf8 never reads more than len bytes from src. If the
sequence is longer than that, or the memory area contains an invalid
sequence, scan_utf8 returns 0 and does not touch dest.
The length of the longest UTF-8 sequence is 5. If the buffer is longer
than that, and scan_utf8 fails, then the data was not a valid UTF-8