From: Tom Lane <tgl@sss.pgh.pa.us>
To: Michael Paquier <michael@paquier.xyz>
cc: Japin Li <japinli@hotmail.com>, amutu@amutu.com,
        pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #18134: ROW_COUNT do not set to 0 when psql's \gset command
 get no rows returned
In-reply-to: <ZRizujD8HMMpSXG5@paquier.xyz>
References: <18134-87126d90cb4dd049@postgresql.org>
 <MEYP282MB1669351F137C61C19FA9749BB6C2A@MEYP282MB1669.AUSP282.PROD.OUTLOOK.COM>
 <ZRizujD8HMMpSXG5@paquier.xyz>
Comments: In-reply-to Michael Paquier <michael@paquier.xyz>
	message dated "Sun, 01 Oct 2023 08:48:10 +0900"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1615011.1696121047.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Sat, 30 Sep 2023 20:44:07 -0400
Message-ID: <1615012.1696121047@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/1615012.1696121047%40sss.pgh.pa.us>
Precedence: bulk

Michael Paquier <michael@paquier.xyz> writes:
> On Wed, Sep 27, 2023 at 06:00:58PM +0800, Japin Li wrote:
>> It seems SaveResultVariables() was lost when executing failed. Attached=
 fix it.

> Yeah, I think that you're right here, the variables should be set.
> SetResultVariables() ought to be called even if \gset failed to return
> a result.

Agreed.  The adjacent comment is now a lie though.  Maybe just

- 		/* set variables on last result if all went well */
+ 		/* set variables from last result */

Bigger-picture question: should we back-patch this, or just fix it in
HEAD?  It's wrong, but I'm not sure it's so obviously wrong that there
could not be somebody depending on the existing behavior.

			regards, tom lane