Re: Oddity with parallel safety test for scan/join target in grouping_planner

Etsuro Fujita <fujita(dot)etsuro(at)lab(dot)ntt(dot)co(dot)jp> writes:
> (2019/02/28 0:52), Robert Haas wrote:
>> On Tue, Feb 26, 2019 at 7:26 AM Etsuro Fujita
>> <fujita(dot)etsuro(at)lab(dot)ntt(dot)co(dot)jp> wrote:
>>> The parallel safety of the final scan/join target is determined from the
>>> grouping target, not that target, which [ is wrong ]

>> Your patch looks right to me.

> I think it would be better for you to take this one. Could you?

I concur with Robert that this is a brown-paper-bag-grade bug in
960df2a97. Please feel free to push (and don't forget to back-patch).

The only really interesting question is whether it's worth adding
a regression test for. I doubt it; the issue seems much too narrow.
Usually the point of a regression test is to prevent re-introduction
of the same/similar bug, but what class of bugs would you argue
we'd be protecting against?

regards, tom lane