Re: [HACKERS] Surjective functional indexes

Greetings,

* Andres Freund (andres(at)anarazel(dot)de) wrote:
> On 2019-01-14 18:46:18 -0500, Stephen Frost wrote:
> > * Tom Lane (tgl(at)sss(dot)pgh(dot)pa(dot)us) wrote:
> > > Andres Freund <andres(at)anarazel(dot)de> writes:
> > > > On 2019-01-14 18:03:24 -0500, Tom Lane wrote:
> > > >> Do we want to revert entirely, or leave the "recheck_on_update" option
> > > >> present but nonfunctional?
> > >
> > > > I think it depends a bit on whether we want to revert in master or
> > > > master and 11. If only master, I don't see much point in leaving the
> > > > option around. If both, I think we should (need to?) leave it around in
> > > > 11 only.
> > >
> > > After a few minutes' more thought, I think that the most attractive
> > > option is to leave v11 alone and do a full revert in HEAD. In this
> > > way, if anyone's attached "recheck_on_update" options to their indexes,
> > > it'll continue to work^H^H^H^Hdo nothing in v11, though they won't be
> > > able to migrate to v12 till they remove the options. That way we
> > > aren't bound to the questionable design and naming of that storage
> > > option if/when we try this again.
> >
> > So the plan is to add a check into pg_upgrade to complain if it comes
> > across any cases where recheck_on_update is set during its pre-flight
> > checks..?
>
> I don't think so.

I could possibly see us just ignoring the option in pg_dump, but that
goes against what Tom was suggesting, since users wouldn't see an error
if we don't dump the option out..

> > Or are you suggesting that pg_dump in v12+ would throw errors if it
> > finds that set? Or that we'll dump it, but fail to allow it into a
> > v12+ database? What if v12 sees "recheck_on_update='false'", as a v11
> > pg_dump might output today?
>
> It'll just error out on restore (including the internal one by
> pg_upgrade). I don't see much point in doing more, this isn't a widely
> used option by any stretch.

This.. doesn't actually make sense. The pg_upgrade will use v12+
pg_dump. You're saying that the v12+ pg_dump will dump out the option,
but then the restore will fail to understand it?

Thanks!

Stephen