Previously the test cases relied on ASCII coding for characters, and didn't test incremental byte by byte checksum calculations.