From: Tom Lane <tgl@sss.pgh.pa.us>
To: Vinay Banakar <vinay.s.banakar@gmail.com>
cc: Andres Freund <andres@anarazel.de>, pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #16006: Update queries fail on a table having any policy with
 a function that takes a whole-row var as arg
In-reply-to: 
 <CAE_mzYQuB+NEN8ja+UwUUqpcAUm3q+84cDQYg=U77daLuNOX8w@mail.gmail.com>
References: <16006-99290d2e4642cbd5@postgresql.org>
 <28076.1568320435@sss.pgh.pa.us>
 <20190912211047.kr7e4pn7iof76jfx@alap3.anarazel.de>
 <1141.1568327305@sss.pgh.pa.us>
 <CAE_mzYQuB+NEN8ja+UwUUqpcAUm3q+84cDQYg=U77daLuNOX8w@mail.gmail.com>
Comments: In-reply-to Vinay Banakar <vinay.s.banakar@gmail.com>
	message dated "Fri, 13 Sep 2019 09:43:21 +0530"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <23377.1568348957.1@sss.pgh.pa.us>
Date: Fri, 13 Sep 2019 00:29:17 -0400
Message-ID: <23378.1568348957@sss.pgh.pa.us>
Precedence: bulk

Vinay Banakar <vinay.s.banakar@gmail.com> writes:
> Thank you for finding the root cause, I just checked on 9.15.17 and it is
> still failing here also.
> But I distinctly remember it working earlier, not sure which version of pg
> (at least <9.5.16) though as I have lost that earlier setup.

Oh!  I misread your report as being that it worked on 9.5.x and failed
on 9.6.x.  If it really did work on some earlier 9.5.x version, then
we must've busted it with a patch ... but what?  That bizarre code in
ExecEvalWholeRowVar is much older than 9.5 (looks like it dates to
8e617e29a), and the mistake proper goes back to the introduction of
WITH CHECK OPTION in 9.4 (4cbe3ac3e).  I'm not seeing any other
moving parts here.

			regards, tom lane