Re: Oddity in tuple routing for foreign partitions

At Thu, 26 Apr 2018 21:16:38 +0900, Etsuro Fujita <fujita(dot)etsuro(at)lab(dot)ntt(dot)co(dot)jp> wrote in <5AE1C326(dot)6040201(at)lab(dot)ntt(dot)co(dot)jp>
> (2018/04/26 20:06), Kyotaro HORIGUCHI wrote:
> > Please rewrite it to use not array reference, but pointer
> > reference if one mtstate logically holds just one resultRelInfo.
>
> Maybe I don't understand your words correctky, but in that UPDATE
> case, I think that mtstate can have multiple ResultRelInfo.

Ah, mtstate has the same number of resultRelInfo with subplans
and it is *written* in the comment just above:( And it is exactly
for the UPDATE case. Sorry for the silly comment.

> >>>> Anyway, I think that
> >>>> the former is more like an improvement rather than a fix, so it would
> >>>> be
> >>>> better to leave that for another patch for PG12?
> >>>
> >>> I agree, so I'm dropping the patch for 1.
> >>
> >> OK, let's focus on #2!
> >>
> >>> See attached an updated version with changes as described above.
> >>
> >> Looks good to me. Thanks for the updated version!
> >
> > Agreed on all points above.
>
> Thanks for reviewing!

I'm happy if it helps you.

regards.

--
Kyotaro Horiguchi
NTT Open Source Software Center