On Fri, Jul 25, 2025 at 08:57:29AM +0000, Bertrand Drouvot wrote:
> This was not needed before fc415edf8ca8, and I think it was better to use
> pgstat_have_pending_wal() to not have to set a flag to every places pgWalUsage.XX
> changes.
At the end, this cost makes the separation of the code layers a bit
better, by having pgstat.c not know about the specifics around the
fixed-numbered stats, so..
> Having said that, I think the single global flag patch is pretty straightforward
> and LGTM.
I have used that and applied it down to v18, closing the open item.
--
Michael