>> better yet: create generic byte-in-Nbyteinteger swap functions.
>> Then define ntoh* and hton* to use them, if appropriate for the
>> port.

> Hear.  Hear.

Me too.  (No, I'm not sending this from AOL! :-)

> Think of MD5 needing to byte swap on big-endian platforms.

...huh?  When I did my MD5 stuff, I didn't see anything that needed
byte-swapping on big-endian hardware.  Closest thing I see is that on a
little-endian machine some of the buffer-filling code amounts to a
bcopy.  Not something I would recommend relying on; there's no need to
optimize to that extent - and indeed, my MD5 code doesn't care whether
it's on big-endian or little-endian or even PDP-endian hardware.  As
long as it has some unsigned integral datatype of at least 32 bits (and
8-bit unsigned chars, actually), it should be fine.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B