Why are both little and big endian crc32 versions created?
Surely it should be enough to detect endian at compile
time and only compile the one needed for that particular CPU?
Now one wastes 16 KB in tables and the duplicated code that goes with it.
Jocke