From: Tom Lane <tgl@sss.pgh.pa.us>
To: Dean Rasheed <dean.a.rasheed@gmail.com>
cc: Andres Freund <andres@anarazel.de>,
        Tomas Vondra <tomas.vondra@2ndquadrant.com>,
        PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Subject: Re: Multivariate MCV stats can leak data to unprivileged users
In-reply-to: 
 <CAEZATCV5fo2E6K5EVN2uytGjMvhTKTizP_rOW67nPqNK8jHAAg@mail.gmail.com>
References: 
 <CAEZATCUhT9rt7Ui=Vdx4N==VV5XOK5dsXfnGgVOz_JhAicB=ZA@mail.gmail.com>
 <20190513223605.lwj65akict3tpetv@development>
 <CAEZATCWFCN2yBpqpw4kw6xhwdPNz8DERPOWNHHQ_qHV4=tbJ8A@mail.gmail.com>
 <20190517202938.lkx6q2u6woiptrsa@alap3.anarazel.de>
 <CAEZATCX9yU2eKL7-dCWTv1pMMi1hb+PsZyM6gjBxupDnMuEzRw@mail.gmail.com>
 <CAEZATCV5fo2E6K5EVN2uytGjMvhTKTizP_rOW67nPqNK8jHAAg@mail.gmail.com>
Comments: In-reply-to Dean Rasheed <dean.a.rasheed@gmail.com>
	message dated "Sat, 18 May 2019 11:48:35 +0100"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <21268.1558192388.1@sss.pgh.pa.us>
Date: Sat, 18 May 2019 11:13:08 -0400
Message-ID: <21269.1558192388@sss.pgh.pa.us>
Precedence: bulk

Dean Rasheed <dean.a.rasheed@gmail.com> writes:
> On the other hand, pg_dump relies on pg_statistic_ext to work out
> which extended statistics objects to dump. If we were to change that
> to use pg_stats_ext, then a user dumping a table with RLS using the
> --enable-row-security flag wouldn't get any extended statistics
> objects, which would be a somewhat surprising result.

It seems like what we need here is to have a separation between the
*definition* of a stats object (which is what pg_dump needs access
to) and the current actual *data* in it.  I'd have expected that
keeping those in separate catalogs would be the thing to do, though
perhaps it's too late for that.

			regards, tom lane