We currently copy a byte at a time when reading from an input from LibFuzzer. We should optimise for the byte aligned case by using `memcpy()`.