Date: Sun, 3 Aug 2025 17:03:21 -0700
From: Noah Misch <noah@leadboat.com>
To: Robert Haas <robertmhaas@gmail.com>
Cc: pgsql-hackers@postgresql.org
Subject: Re: Test instability when pg_dump orders by OID
Message-ID: <20250804000321.8b.nmisch@google.com>
References: <20250707192654.9e.nmisch@google.com>
 <CA+TgmoYnAiypU5sYstJ0nCqr+VVubf5wczANg=KqEeOYmdVeAw@mail.gmail.com>
 <20250718191702.45.nmisch@google.com>
 <CA+Tgmoa8FumzaUUSnjK0x=6_oXDXggQPpzgPfX-KE=PVrTyamg@mail.gmail.com>
 <20250725022740.30.nmisch@google.com>
 <CA+TgmoZXmAZmB9OZaXTiy-HG2_1d7UiYrFKXf17C_yLfPa5pQg@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: 
 <CA+TgmoZXmAZmB9OZaXTiy-HG2_1d7UiYrFKXf17C_yLfPa5pQg@mail.gmail.com>
User-Agent: Mutt/2.2.12 (2023-09-09)
Archived-At: 
 <https://www.postgresql.org/message-id/20250804000321.8b.nmisch%40google.com>
Precedence: bulk

On Fri, Jul 25, 2025 at 02:01:01PM -0400, Robert Haas wrote:
> On Thu, Jul 24, 2025 at 10:27 PM Noah Misch <noah@leadboat.com> wrote:
> > I regret missing those in v1.  I've attached v2, including branch-specific
> > patches.  I'll first need to back-patch 350e6b8, which fixed sorting of CREATE
> > RULE, to v17 and earlier.  Since 350e6b8 is conflict-free all the way back to
> > v13, I'm not attaching it.
> 
> Back-patching 350e6b8 to facilitate back-patching this seems OK. I did
> a read-through of dobjcomp20-disambiguate-v2.patch and have no further
> review comments. I did not read through the back-patched versions on
> the assumption that back-porting was straightforward enough that a
> separate review was not required.

Pushed as 0decd5e.  v14 supports binary-upgrade from v8.3 and regular dump
from v8.0.  That required two other changes.  First, pg_opclass.opcmethod had
name opcamid until v8.3 (commit a78fcfb).  Accounting for both names was
trivial.  Second, pg_am first had fixed OID AccessMethodRelationId in v8.1
(commit 7c13781).  The find*ByOid functions have been assuming fixed catalog
OIDs since v15's commit 92316a4.  The $SUBJECT commit added
findAccessMethodByOid() to all branches, so I changed the v14/v13
findAccessMethodByOid() to be more like v14/v13 findOprByOid(), which doesn't
assume AccessMethodRelationId.  If folks want more details, let me know.