Using a filesize in bytes as offset into a Python string is not a good idea if said file contains non-ASCII characters.