in reply to RE (tilly) 3: sort performancein thread sort performance
*sigh*
(Stupid user tricks.)
I get "schartz" 5%-30% faster than "straight" with "tye" (which you didn't time) being 50%-100% faster still (all using "cmp"). (: