Re: path toward faster partition pruning

On 2017/09/27 1:51, Robert Haas wrote:
> On Tue, Sep 26, 2017 at 10:57 AM, Jesper Pedersen
> <jesper(dot)pedersen(at)redhat(dot)com> wrote:
>> One could advocate (*cough*) that the hash partition patch [1] should be
>> merged first in order to find other instances of where other CommitFest
>> entries doesn't account for hash partitions at the moment in their method
>> signatures; Beena noted something similar in [2]. I know that you said
>> otherwise [3], but this is CommitFest 1, so there is time for a revert
>> later, and hash partitions are already useful in internal testing.
>
> Well, that's a fair point. I was assuming that committing things in
> that order would cause me to win the "least popular committer" award
> at least for that day, but maybe not. It's certainly not ideal to
> have to juggle that patch along and keep rebasing it over other
> changes when it's basically done, and just waiting on other
> improvements to land. Anybody else wish to express an opinion?

FWIW, I tend to agree that it would be nice to get the hash partitioning
patch in, even with old constraint exclusion based partition-pruning not
working for hash partitions. That way, it might be more clear what we
need to do in the partition-pruning patches to account for hash partitions.

Thanks,
Amit