Re: Re: qsort fix

Jim Winstead wrote: > Sounds like a good reason to me. (I knew there was on there somewhere.) > > And for something like shuffle(), it's arguably a good thing if it > is worse than O(n) because it will shuffle it that much more. :) I agree, although I'm a little upset with sun on this one... I can't think of many times when I would find it preferable for my program to core dump rather than return an unsorted array... BTW, I did some before and after benchmarks on my home machine, timing the shuffle using the current shuffle with qsort vs a shuffle algorithm. See http://zeus1.tzo.com/diff.php3. Qsort is a bit quicker on smaller arrays, but with such small times, it's pretty irrelevant. -- Chad Cunningham ccunning@math.ohio-state.edu