From: Tom Lane <tgl@sss.pgh.pa.us>
To: bernice.southey@gmail.com
cc: pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #19072: New-style SQL language function referencing a temp
 table behaves oddly
In-reply-to: <19072-819bab455efebdd7@postgresql.org>
References: <19072-819bab455efebdd7@postgresql.org>
Comments: In-reply-to PG Bug reporting form <noreply@postgresql.org>
	message dated "Sat, 04 Oct 2025 16:51:33 -0000"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1526597.1759598760.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Sat, 04 Oct 2025 13:26:00 -0400
Message-ID: <1526598.1759598760@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/1526598.1759598760%40sss.pgh.pa.us>
Precedence: bulk

PG Bug reporting form <noreply@postgresql.org> writes:
> When created in psql, a new-style SQL language function referencing a te=
mp
> table disappears with the session, as if it were a temporary function.

Yeah, this is expected, since the parser will create a pg_depend
linkage from the function to the temp table.

> When created in pgAdmin, the function survives.

Really?  I don't use pgAdmin, but I don't see how it could affect the
object dependency rules.  Perhaps it's not closing the originating
session when you think it is?

There is a nearby thread proposing forbidding new-style functions
from having dependencies on temp objects [1].  Curious to know if
you think that'd be a good answer.

			regards, tom lane

[1] https://www.postgresql.org/message-id/flat/19cf6ae1-04cd-422c-a760-d7e=
75fe6cba9%40uni-muenster.de