On Tue, Jan 25, 2022 at 07:30:33PM +0000, Bossart, Nathan wrote:
> I think the patch is in decent shape. There may be a few remaining
> places where GetMaxBackends() is called repeatedly in the same
> function, but IIRC v4 already clears up the obvious ones. I don't
> know if this is worth worrying about too much, but I can create a new
> version if you think it is important.
There are such cases in FindLockCycleRecurse(), GetLockConflicts(),
GetLockStatusData() and InitProcGlobal(), as far as I can see.
Hmm. I have been looking at this patch, and the lack of centralized
solution that could be used for other GUCs worries me like Fujii-san,
even if this would prevent an incorrect use of MaxBackends in contexts
where it should not be used because it is not initialized yet. I
don't think it is a good idea in the long-term to apply this as-is.
--
Michael