| From: | David Rowley <dgrowleyml(at)gmail(dot)com> |
|---|---|
| To: | Jaime Casanova <jcasanov(at)systemguards(dot)com(dot)ec>, Andres Freund <andres(at)anarazel(dot)de> |
| Cc: | Yura Sokolov <y(dot)sokolov(at)postgrespro(dot)ru>, PostgreSQL Developers <pgsql-hackers(at)lists(dot)postgresql(dot)org>, David Rowley <dgrowley(at)gmail(dot)com> |
| Subject: | Re: Use simplehash.h instead of dynahash in SMgr |
| Date: | 2021-09-27 03:30:25 |
| Message-ID: | CAApHDvqQXsbNZ=pZcDuSw3vdgr6+n5QVHQFQHmF3LyUXpedbeQ@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Fri, 24 Sept 2021 at 20:26, Jaime Casanova
<jcasanov(at)systemguards(dot)com(dot)ec> wrote:
> Are you planning to work on this in this CF?
> This is marked as "Ready for committer" but it doesn't apply anymore.
I've attached an updated patch. Since this patch is pretty different
from the one that was marked as ready for committer, I'll move this to
needs review.
However, I'm a bit disinclined to go ahead with this patch at all.
Thomas made it quite clear it's not for the patch, and on discussing
the patch with Andres, it turned out he does not like the idea either.
Andres' argument was along the lines of bitmaps being slow. The hash
table uses bitmaps to record which items in each segment are in use. I
don't really agree with him about that, so we'd likely need some more
comments to help reach a consensus about if we want this or not.
Maybe Andres has more comments, so I've included him here.
David
| Attachment | Content-Type | Size |
|---|---|---|
| v2-0001-Use-densehash.h-hashtables-in-SMgr.patch | application/octet-stream | 46.0 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Amit Kapila | 2021-09-27 03:45:53 | Re: Skipping logical replication transactions on subscriber side |
| Previous Message | Amit Kapila | 2021-09-27 03:24:26 | Re: Skipping logical replication transactions on subscriber side |