@dmsnell significantly improved the speed of the UTF-8 decoder, let's bring those improvements into this project.