Re: [boost] Strange use of random values in regression test ...