At Fri, 01 Apr 2022 14:44:30 +0900 (JST), Kyotaro Horiguchi <horikyota.ntt@gmail.com> wrote in
> At Sun, 13 Feb 2022 17:35:38 +0300, Yura Sokolov <y.sokolov@postgrespro.ru> wrote in
> > And there are checks for `takenDuringRecovery` in `heapgetpage` and
> > `heapam_scan_sample_next_tuple`. Are this checks affected by the change?
> > Neither the preceding discussion nor commit message answer me.
>
> The snapshot works correctly, but for the heapgetpage case, it foces
> all_visible to be false. That unnecessarily prevents visibility check
> from skipping.
>
> An annoying thing in SnapBuildInitialSnapshot is that we don't know
> the number of xids before looping over the xid range, and we don't
> want to bother sorting top-level xids and subxids unless we have to do
> so.
>
> Is it better that we hassle in SnapBuildInitialSnapshot to create a
> !takenDuringRecovery snapshot?
So this is that. v5 creates a regular snapshot.
By the way, is there any chance this could be committed to 15?
Otherwise I'll immediately move this to the next CF.
regards.
--
Kyotaro Horiguchi
NTT Open Source Software Center