hi Chris,

The KNF reference wasn't so much a strike against the code Itojun
posted; it was a requirement on any substitute.


One might argue that (to quote from the code Steve Rumble mentioned)
"ugliness and redundancy" violate KNF, but that wasn't my intent.
leaving aside the K&R definitions as I honestly dont remember whether
new code is allowed/encouraged/supposed to use ANSI, or not.

Personally, the side-effect and break in that `copy as many bytes as
will fit' loop strikes as just atrocious. Of course ymmv.