How can I get the byte sequence of a particular UTF-8 string? I'm seeing what looks like some bug in the regex engine which is only triggered in some edge cases, and I'd like to know exactly what data it's working on.