Normally, UTF-8 works with C strings no problem, but it's possible that a multi-byte UTF-8 encoded codepoint is only halfway copied when it reaches the capacity and throws the NULL at the end.