Message-ID: <44220da166d8f1ef74b2bf76c60a7663dcf9dc07.camel@j-davis.com>
Subject: Re: Avoid orphaned objects dependencies, take 3
From: Jeff Davis <pgsql@j-davis.com>
To: Bertrand Drouvot <bertranddrouvot.pg@gmail.com>, Heikki Linnakangas
	 <hlinnaka@iki.fi>
Cc: Robert Haas <robertmhaas@gmail.com>, Roman Eskin <r.eskin@arenadata.io>,
  Michael Paquier <michael@paquier.xyz>, Alexander Lakhin
 <exclusion@gmail.com>, pgsql-hackers@lists.postgresql.org,  Tom Lane
 <tgl@sss.pgh.pa.us>
Date: Thu, 04 Jun 2026 10:17:57 -0700
In-Reply-To: <5315d15a42109297259d1a3264ad09e363eb98df.camel@j-davis.com>
References: 
	<CA+TgmoZh8yXoQ2AF-VFSTswhin0o+BZ78AaOsWZskCW1GHBd6g@mail.gmail.com>
	 <aEJ+oNANWutG8aDv@ip-10-97-1-34.eu-west-3.compute.internal>
	 <9f5f0dd4-96d5-4d90-9436-62d96e0298f9@arenadata.io>
	 <ad/awXLnMuc6Rf5M@bdtpg> <afCXPVNuvpf5ECGZ@bdtpg>
	 <CA+TgmobtzkG9HzGtDXbgb3KVE46qHuSvGcuKHgAGQ1aejCoLLw@mail.gmail.com>
	 <agsCqlLTytZCudMv@bdtpg> <aebf583b-b904-4035-a1c1-a4c5ad064dec@iki.fi>
	 <ahYAKgbx05IQdjwR@bdtpg> <3963e573-4852-4d25-b741-7c5c7e1f9140@iki.fi>
	 <ah1PIyqtT+427dWb@bdtpg>
	 <5315d15a42109297259d1a3264ad09e363eb98df.camel@j-davis.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.52.3-0ubuntu1.1 
MIME-Version: 1.0
Archived-At: 
 <https://www.postgresql.org/message-id/44220da166d8f1ef74b2bf76c60a7663dcf9dc07.camel%40j-davis.com>
Precedence: bulk

On Wed, 2026-06-03 at 11:08 -0700, Jeff Davis wrote:
> RangeVarGetRelidExtended() coordinates three things:
>=20
> =C2=A0 - name lookup
> =C2=A0 - lock
> =C2=A0 - ACL check
>=20
> whereas recheckAclAndLock() only coordinates the latter two.

Given that we don't do another name lookup, the object Oid doesn't
change, and it's not obvious why we need a loop in this path.

A sequence like:

Earlier during DDL processing:
  0. Name lookup and ACL check (and track ACLs)

When recording dependencies:
  1. Lock object
  2. Check that it still exists, error if not
  3. recheck tracked ACLs, error if failure

could work too, right?

I see why you might want to do the checks while not holding the lock,
but it doesn't seem like a requirement (if the user doesn't have
permissions it should fail quickly and release). In any case, it's
worth a comment explaining why the loop is there.

Regards,
	Jeff Davis