Re: Heap's backwards scan scans the incorrect pages with heap_setscanlimits() - Mailing list pgsql-hackers

From David Rowley
Subject Re: Heap's backwards scan scans the incorrect pages with heap_setscanlimits()
Date
Msg-id CAApHDvpP9RAFz3bmgS_kTnmbADBm=JWT5jG4LTsenZ=XXVzVUA@mail.gmail.com
Whole thread Raw
In response to Heap's backwards scan scans the incorrect pages with heap_setscanlimits()  (David Rowley <dgrowleyml@gmail.com>)
List pgsql-hackers
On Thu, 21 Jan 2021 at 13:16, David Rowley <dgrowleyml@gmail.com> wrote:
> Proposed patch attached.

I ended up pushing a slightly revised version of this which kept the
code the same as before when rs_numblocks had not been changed. I
backpatched to 9.5 as it seemed low risk and worthy of stopping some
head-scratching and a future report for any extension authors that
make use of heap_setscanlimits() with backwards scans at some point in
the future.

David



pgsql-hackers by date:

Previous
From: "Joel Jacobson"
Date:
Subject: The mysterious pg_proc.protrftypes
Next
From: "tsunakawa.takay@fujitsu.com"
Date:
Subject: RE: libpq debug log