Message-ID: <ab67c99bfb5dea7bae18c77f96442820d19b5448.camel@cybertec.at>
Subject: Re: Fix output of zero privileges in psql
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: "David G. Johnston" <david.g.johnston@gmail.com>
Cc: Erik Wienhold <ewie@ewie.name>, PostgreSQL Hackers
	 <pgsql-hackers@lists.postgresql.org>
Date: Tue, 17 Oct 2023 15:05:56 +0200
In-Reply-To: <CAKFQuwbWt6pvSFERENDh-v8bsahgSd1NDfxsRfFT4g_MNDu+ZQ@mail.gmail.com>
References: <1966228777.127452.1694979110595@office.mailbox.org>
	 <84f7a1dd3052544974d5b4d5df7beb5a244b48fa.camel@cybertec.at>
	 <4tgcecleqgkynyat7t3innrsdpopeiukumb7s4f2gdw3zuzidb@27urnfsqebdy>
	 <b65d6d7a6ba0ffb13cfa82ba694779fa131f9c64.camel@cybertec.at>
	 <hltc5k6wgxkrrsd5wuzt45elxr6wiyvgqkixmfqfkoxirpplsu@zt4uhz5yhdy4>
	 <30ef7918c8f1a0b811d11350e7f658a3c3439441.camel@cybertec.at>
	 <lir6jx6hjhsqkskekrhtoflbjbedv7wqowvz5wc2a742by5i55@uqeiwszaksvs>
	 <b210597b476edb530531f03adcf881d1e84fe086.camel@cybertec.at>
	 <5uytof65cguide4mjhxzakd5w74hpuppjsrqfwuywtcf2r2xwl@lbi4fc374nig>
	 <28a1d403ad2499865a5da1be9dc719b1218f8a8f.camel@cybertec.at>
	 <cww6udzxvxorpznocvufq42mly3qyi3wcjx45nhlbj3wm6rjfc@36pzmsfodow5>
	 <dea834dc421ae8fa60fbe3280e364cd5a9a278d7.camel@cybertec.at>
	 <CAKFQuwbWt6pvSFERENDh-v8bsahgSd1NDfxsRfFT4g_MNDu+ZQ@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.48.4 (3.48.4-1.fc38) 
MIME-Version: 1.0
Archived-At: <https://www.postgresql.org/message-id/ab67c99bfb5dea7bae18c77f96442820d19b5448.camel%40cybertec.at>
Precedence: bulk

On Mon, 2023-10-16 at 19:05 -0700, David G. Johnston wrote:
> Reading both threads I'm not seeing any specific rejection of the
> solution that we simply represent empty privileges as "(none)".
>=20
> I see an apparent consensus that if we do continue to represent it
> as NULL that the=C2=A0printout=C2=A0should=C2=A0respect \pset null.
>=20
> Tom commented in favor of (none) on the other thread with some
> commentary regarding how extremely rare it is; then turns around
> and is "fairly concerned" about changing the current blank presentation
> of its value which is going to happen for some people regardless
> of which approach is chosen.
>=20
> Stuart, the original user complainant, seems to favor (none)
>=20
> Erik seems to favors (none)
>=20
> I favor (none)
>=20
> It's unclear to me whether you=C2=A0Laurenz are against (none) or just
> trying to go with the group consensus that doesn't actually seem to
> be against (none).
>=20
> I'm in favor of iterating on v1 on this thread (haven't read it yet)
> and presenting it as the proposed solution.=C2=A0 If that ends up getting
> shot down we can revive v5 (still need to review as well).

Thanks for that summary.  I prefer my version (simply display NULLs
as NULLs), but I am not wedded to it.  I had the impression that Tom
would prefer that too, but is woried if the incompatibility introduced
would outweigh the small benefit (of either patch).

So it is clear that we don't have a consensus.

I don't think I want to go ahead with my version of the patch unless
there is more support for it.  I can review Erik's original code, if
that design meets with more favor.

> We should probably post on that thread that this one exists and post a li=
nk to it.

Perhaps a good idea, yes.

Yours,
Laurenz Albe