> (Aside: presumably we could walk thru the string destructively,> in-place, swapping bytes; I think that would be theoretically> quickest...)
Hmmm... I guess it will not work für UTF-8 or any other multibyte charset