Re: fibnum3

[Paul Vixie]

noting that we can't get rid of the conditional branch no matter whether we 
assign twice or assign once, regardless of whether the condition is expressed 
as a trinary expression or an if-else chain. i didn't mean to imply otherwise.