MIME-Version: 1.0
References: <fa7903f6-b050-4c62-aa52-6f6d5d345770@dunslane.net>
 <CAApHDvr36T+9XZYTRkHkGm82Wo5YY6wXwkRDceu2-cyGEvHzjw@mail.gmail.com>
 <50d3a76e-afaa-44ed-aef5-6fb7f23cccab@dunslane.net> <5bef8bb0-7a50-4bcf-b052-2a12c3cda0f5@joeconway.com>
 <CA+Tgmob78LN+Tkgms=dQyBsLyGWBuehv7_kEvBfjwmj6fDsr7g@mail.gmail.com>
 <3416373.1722449026@sss.pgh.pa.us> <70b9e758-1050-4365-9d67-cc826faaf6a9@joeconway.com>
 <13ec5d8fc5e3c58053a8762634a4921e3387f57d.camel@cybertec.at>
 <e42ce45b-e9c3-4680-bbc4-b59c32cfc7d0@joeconway.com> <3737863.1722522365@sss.pgh.pa.us>
In-Reply-To: <3737863.1722522365@sss.pgh.pa.us>
From: Jacob Champion <jacob.champion@enterprisedb.com>
Date: Thu, 1 Aug 2024 13:51:41 -0700
Message-ID: <CAOYmi+=_VU9MLd_BknwuEEoPBVwDRByPCAc0dW+S9XFFe7vACw@mail.gmail.com>
Subject: Re: can we mark upper/lower/textlike functions leakproof?
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Joe Conway <mail@joeconway.com>, Laurenz Albe <laurenz.albe@cybertec.at>, 
	Robert Haas <robertmhaas@gmail.com>, Andrew Dunstan <andrew@dunslane.net>, 
	David Rowley <dgrowleyml@gmail.com>, 
	PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: <https://www.postgresql.org/message-id/CAOYmi%2B%3D_VU9MLd_BknwuEEoPBVwDRByPCAc0dW%2BS9XFFe7vACw%40mail.gmail.com>
Precedence: bulk

On Thu, Aug 1, 2024 at 7:26=E2=80=AFAM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Are you proposing that we invent two levels of leakproofness
> with different guarantees?  That seems like a mess, not least
> because there are going to be varying opinions about where we
> should set the bar for the lower level.

It kind of reminds me of the kernel's "paranoia level" tunables, which
seem to proliferate in weird ways [1] and not make for a particularly
great UX.

--Jacob

[1] https://askubuntu.com/questions/1400874/what-does-perf-paranoia-level-f=
our-do