From: Tom Lane <tgl@sss.pgh.pa.us>
To: Peter Eisentraut <peter@eisentraut.org>
cc: Alexander Lakhin <exclusion@gmail.com>,
        Robert Haas <robertmhaas@gmail.com>,
        PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: TupleDescAttr bounds checks
In-reply-to: <55199b03-dd62-46d3-be0a-e39217b27333@eisentraut.org>
References: 
 <CA+TgmoacixUZVvi00hOjk_d9B4iYKswWP1gNqQ8Vfray-AcOCA@mail.gmail.com>
 <1628959.1774023745@sss.pgh.pa.us>
 <CA+TgmoaiPqnV2ePY-DyQUVGnK5sJ5ryPCDYpn_+1Ua4JwfoqRw@mail.gmail.com>
 <1631625.1774025167@sss.pgh.pa.us> <3429884.1774238081@sss.pgh.pa.us>
 <CA+TgmoZgc1m5Z7RFf9U7s_DgXE8GqnOB95fEWurdKFMR_U2xpQ@mail.gmail.com>
 <6f435023-8ab6-47c2-ba07-035d0c4212f9@gmail.com>
 <3956199.1775313002@sss.pgh.pa.us> <3956971.1775313533@sss.pgh.pa.us>
 <4035884.1775317104@sss.pgh.pa.us>
 <55199b03-dd62-46d3-be0a-e39217b27333@eisentraut.org>
Comments: In-reply-to Peter Eisentraut <peter@eisentraut.org>
	message dated "Sat, 04 Apr 2026 23:52:02 +0200"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <504240.1775360727.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Sat, 04 Apr 2026 23:45:27 -0400
Message-ID: <504241.1775360727@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/504241.1775360727%40sss.pgh.pa.us>
Precedence: bulk

Peter Eisentraut <peter@eisentraut.org> writes:
> On 04.04.26 17:38, Tom Lane wrote:
>> After poking at that: testing tableoid does sort of work, in that it
>> reads as the OID of the target table named in COPY.  But I think any
>> rational use for a test on tableoid here would be in connection with
>> a partitioned target table, and the user would wish it to read as the
>> OID of the destination partition.  So I think we should disallow
>> tableoid along with the other system columns, pending somebody having
>> the ambition to make that work.

> I think this is the same issue that was discussed here:
> https://www.postgresql.org/message-id/flat/30c39ee8-bb11-4b8f-9697-45f7e=
018a8d3%40eisentraut.org
> There was no conclusion there, but I agree with the proposal to prohibit=
 =

> this use.

Ah, indeed.  Jian's patch in that thread seems rough but potentially
workable to me, but seemingly the thread tailed off for lack of
interest.  I don't want to revive it now as part of a bug fix.

Disallowing tableoid for now, and then re-allowing it if someone
picks up that patch down the road, seems like a good solution.
For one thing, since that patch changes the semantics of tableoid
in COPY WHERE, I think it'd be a good idea to have a release or
two in between where we throw error.  That'd be helpful to flush
out any field usages that might be affected.

			regards, tom lane