Re: Proposal: Conflict log history table for Logical Replication

Hi,

On Thu, Nov 5, 2026 at 04:54 PM shveta malik wrote:
>
> The xid fix looks okay to me. One minor point: in
> CheckAndReportConflict(), the epoch-based conversion

Thanks for reviewing.

>
> conflicttuple->xmin =
> FullTransactionIdFromAllowableAt(ReadNextFullTransactionId(), xmin);
>
> isn’t strictly needed. These are ERROR-level conflicts, which are
> never written to the conflict log table, so conflicttuple->xmin is
> only consumed by the server-log in ReportApplyConflict() after
> truncating back to the 32-bit xid (epoch thrown). But I think, it is
> good to still retain it as-is for consistency with the other
> conflict-detection sites.

I also think it's fine to keep it for now for consistency.

>
> ~~
>
> I am not sure if you have started taking comments on 002 yet. Here are
> a few comments; feel free to defer them to later versions if we want
> to conclude on 001 first.

Thanks. For now, I'll make a note of these and update 0002 later once
0001 is ready.

Here is the new patch version, which addresses Peter's comments on the
GUC lc_monetary. Also adjusted the doc a bit to make it more generic, so
we don't have to keep adding new entries there.

Per the internal review, I've also improved a few code comments and
removed one header inclusion for the new full xid logic.

0003 has been improved a bit per the internal review, thanks to Vignesh
for doing that.

--
Best Regards,
Zhijie Hou