Re: Inconsistency in determining the timestamp of the db statfile.

On Fri, Sep 11, 2020 at 4:53 AM Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:

> On Thu, Sep 10, 2020 at 6:42 PM Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
> wrote:
> >
> > On 2020-Sep-10, Amit Kapila wrote:
> >
> > > On Thu, Sep 10, 2020 at 1:03 PM Magnus Hagander <magnus(at)hagander(dot)net>
> wrote:
> >
> > > The comments already say what you said in the second suggestion:"The
> > > caller must rely on timestamp stored in *ts iff the function returns
> > > true.". Read iff "as if and only if"
> >
> > I think "must" should be "may" there, if we're nitpicking.
> >
>
> Here, we want to say that "caller can rely on *ts only if the function
> returns true". If we replace 'must' with 'may' then it seems to me we
> are trying to say that caller can ignore the timestamp value, if so,
> why at first place caller has called this function.
>

This is true, but that should really be the decision of the caller, not of
the function.

But again, that's extremely nitpicky, so it doesn't really matter :)

And +1 on the push you did, and the decision not to backpatch it since
there haven't been any reports.

--
Magnus Hagander
Me: https://www.hagander.net/ <http://www.hagander.net/>
Work: https://www.redpill-linpro.com/ <http://www.redpill-linpro.com/>