Content-Type: text/plain;
	charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 13.4 \(3608.120.23.2.7\))
Subject: Re: Collation version tracking for macOS
From: Mark Dilger <mark.dilger@enterprisedb.com>
In-Reply-To: <1315087.1654632634@sss.pgh.pa.us>
Date: Wed, 8 Jun 2022 13:44:32 -0700
Cc: Peter Geoghegan <pg@bowt.ie>,
 Robert Haas <robertmhaas@gmail.com>,
 Thomas Munro <thomas.munro@gmail.com>,
 Jeremy Schneider <schneider@ardentperf.com>,
 Peter Eisentraut <peter.eisentraut@enterprisedb.com>,
 pgsql-hackers <pgsql-hackers@postgresql.org>
Content-Transfer-Encoding: quoted-printable
Message-Id: <A0C774E0-55C0-4566-A520-C0ED7D943B97@enterprisedb.com>
References: <381977b1-0898-cb6f-a427-3b5d873e81bd@enterprisedb.com>
 <CA+hUKGKjkL=APm5nd3vGOtMAD_07nyvrbAaGaDduVD-EKM5Drw@mail.gmail.com>
 <ce1174ea-1942-37f0-421e-9ccff400dbcb@enterprisedb.com>
 <231072.1654273317@sss.pgh.pa.us>
 <1874de62-6bec-4bc1-1d14-0a2730b125da@ardentperf.com>
 <CA+hUKGLwT9xBSodtmio-NnGXfkrfVs02O6kg1iwBJSt-XMhW4g@mail.gmail.com>
 <366234.1654289888@sss.pgh.pa.us>
 <CA+TgmoZyjoRMj=m2s2a=Hvxp05aAwectjUeSP0Nf67Jsz8-cEA@mail.gmail.com>
 <CAH2-WzmXFVw6wAtgLo+0ggin===9TOG8OZC9VorHCu9W7y4Kew@mail.gmail.com>
 <1315087.1654632634@sss.pgh.pa.us>
To: Tom Lane <tgl@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/A0C774E0-55C0-4566-A520-C0ED7D943B97%40enterprisedb.com>
Precedence: bulk


> On Jun 7, 2022, at 1:10 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>=20
> This is not the concern that I have.  I agree that if we tell a user
> that collation X changed behavior and he'd better reindex his indexes
> that use collation X, but none of them actually contain any cases that
> changed behavior, that's not a "false positive" --- that's "it's =
cheaper
> to reindex than to try to identify whether there's a problem".

I don't see this problem as limited to indexes, though I do understand =
why that might be the most common place for the problem to manifest =
itself.

As a simple example, text[] constructed using array_agg over sorted data =
can be corrupted by a collation change, and reindex won't fix it.

If we extend the table-AM interface to allow query quals to be pushed =
down to the table-AM, we might develop table-AMs that care about sort =
order, too.

=E2=80=94
Mark Dilger
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company