On Thu, Apr 21, 2016 at 7:13 AM, Peter Geoghegan <pg@heroku.com> wrote:
> On Wed, Apr 20, 2016 at 2:49 PM, Robert Haas <robertmhaas@gmail.com> wrote:
>> Committed. But I think the regression test needs more thought, so I
>> left that out.
>
> It would be nice if there was a fuzz testing infrastructure that
> verified that parallel plans produce the same answer as serial plans.
Results of parallel plans and serial plans could be stored in
temporary tables in the test, then that's a matter of comparing them I
guess. That's largely doable.
--
Michael
