On Wed, Jan 7, 2009 at 4:09 AM, Walter Bright <newshou...@digitalmars.com> wrote: > Ary Borenszweig wrote: >> >> I think Bill speaks about a stable sort. You can have an unstable sort >> algorithm without having explicity a random invocation. Note that he's >> saying "leads to results that differ from run-to-run when there are elements >> with identical keys". > > And I believe that the sort implementation is completely deterministic. If > Bill is getting different results from different runs, something else is > going on. >
Really? Hmm, I wonder what the reason was, then. It did go away when I switched to a different sort routine. Don't have time to investigate right now, unfortunately. But thank you very much for the info that it should be deterministic. --bb