Обсуждение: Re: BUG #15383: Join Filter cost estimation problem in 10.5

Status update for a commitfest entry.

It looks like there was no real progress on this issue since April. I see only an experimental patch. What kind of
reviewdoes it need right now? Do we need more testing or maybe production statistics for such queries?
 

David, are you going to continue working on it?
Can you, please, provide a short summary of the problem and list open questions for reviewers.

The new status of this patch is: Waiting on Author

Anastasia Lubennikova <lubennikovaav@gmail.com> writes:
> Status update for a commitfest entry.
> It looks like there was no real progress on this issue since April. I see only an experimental patch. What kind of
reviewdoes it need right now? Do we need more testing or maybe production statistics for such queries? 

I think it's probably time to close this entry as RWF.  I don't think
either David or I have any fresh ideas about how to get to something
committable.

The last couple of messages were complaining that adding this topic
to the TODO list would be useless ... but now that Naylor is working
on cleaning that up, maybe it will become less useless.

            regards, tom lane

On Fri, Oct 30, 2020 at 9:16 AM Anastasia Lubennikova <lubennikovaav@gmail.com> wrote:

Status update for a commitfest entry.

It looks like there was no real progress on this issue since April. I see only an experimental patch. What kind of review does it need right now? Do we need more testing or maybe production statistics for such queries?

David, are you going to continue working on it?
Can you, please, provide a short summary of the problem and list open questions for reviewers.

The new status of this patch is: Waiting on Author

I agree that updating a wiki seems like an unappealing conclusion to this entry.  I'm on the fence whether we just leave it in the cf and get a periodic nag when it gets bumped.  As we are describing real problems or potential improvements to live portions of the codebase ISTM that adding code comments near to those locations would be warranted.  The commit message can reference this thread.  There seems to already be a convention for annotating code comments in such a way that they can be searched for - which basically is what sticking it in the wiki would accomplish but the searcher would have to look in the codebase and not on a website.  For the target audience of this specific patch that seems quite reasonable.

David J.

Here we are again.
The commitfest is closed now and the discussion haven't moved from where it was a month ago.
I don't see any use in moving it to the next CF as is and I don't want to return it either, as this is clearly a bug.

I think we have a few options:

1. We can add it into TODO until better times.
2. We can write a patch to address this problem in code comments.
3. We can maybe CC this thread to someone and ask for help. Do you know people, who have  expertise in this area?

None of these options is perfect, but the second option will perhaps be a good compromise.
David, can you, please submit a patch?

-- 
Anastasia Lubennikova
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company