On Sun, 2020-12-27 at 02:09 -0500, Tom Lane wrote:
> Jeff Davis <jdavis@postgresql.org> writes:
> > Fix bug #16784 in Disk-based Hash Aggregation.
>
> Buildfarm thinks this new test case is not completely stable.
>
> It appears that only 32-bit machines are failing, and it's
> not totally consistent even for them. Dunno what to make of it.
I think I just need to disable sort. Because there is a () group, we
don't need a Sort in the plan to exercise the bug.
Committed to master. I'll backport if the tests stabilize. I'm on
vacation now and will return on the 5th, so hopefully we're in a
reasonable state until then. If not, I can just revert my fix and deal
with it when I get back.
Regards,
Jeff Davis