Date: Mon, 23 Mar 2026 06:33:33 +0000
To: Tom Lane <tgl@sss.pgh.pa.us>
From: Pierre Ducroquet <p.psql@pinaraf.info>
Cc: Tomas Vondra <tomas@vondra.me>, Jelte Fennema-Nio <postgres@jeltef.nl>,
 Andres Freund <andres@anarazel.de>, Greg Sabino Mullane <htamfids@gmail.com>,
 Adrien Nayrat <adrien.nayrat@anayrat.info>, Christoph Berg <myon@debian.org>,
 Euler Taveira <euler@eulerto.com>,
 =?utf-8?Q?=C3=81lvaro_Herrera?= <alvherre@kurilemu.de>,
 PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>,
 Andreas Karlsson <andreas@proxel.se>,
 Anthonin Bonnefoy <anthonin.bonnefoy@datadoghq.com>,
 pgsql-hackers <pgsql-hackers@postgresql.org>, Michael Banck <mbanck@gmx.net>
Subject: Re: Change default of jit to off
Message-ID: 
 <gPNlX1QTDtJW2Cyi3fA2OGFvoIybkehUYKCHCe2r3Ujm_0y9V1T13CPI2MA6RyPhLi8zNwNczq1DNrbh1l47eVn6tOwz-qzG2yXw_XtptTQ=@pinaraf.info>
In-Reply-To: <1629115.1774023867@sss.pgh.pa.us>
References: <aYHN-fxswaIgfI-B@msg.df7cb.de>
 <49490d8d-e8b9-497b-9a58-712daadb9ca9@anayrat.info>
 <CAKAnmmKZ_Yg2CEt9Wa-QZWZ7UePfWB9qnk=qDiXZcqNS5MdrGQ@mail.gmail.com>
 <zavalseit5sbjf5gx352ioobtlqtycrs5yczynwzckacdwqh23@mjuvibp2z44t>
 <CAGECzQQdO5OGnxe7bpN1usp+ku+tLBW=SwgLbrULVcFAjH5_yA@mail.gmail.com>
 <5b5d04c3-e34e-4ae9-8b39-12b52e63154a@vondra.me>
 <1629115.1774023867@sss.pgh.pa.us>
Feedback-ID: 175348558:user:proton
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/gPNlX1QTDtJW2Cyi3fA2OGFvoIybkehUYKCHCe2r3Ujm_0y9V1T13CPI2MA6RyPhLi8zNwNczq1DNrbh1l47eVn6tOwz-qzG2yXw_XtptTQ%3D%40pinaraf.info>
Precedence: bulk

Le vendredi 20 mars 2026 =C3=A0 5:25 PM, Tom Lane <tgl@sss.pgh.pa.us> a =
=C3=A9crit=C2=A0:

> Tomas Vondra <tomas@vondra.me> writes:
> > ISTM there's a clear consensus to get this committed for PG19, so
> > barring objections I'll take care of that in the next couple days.
> > Unless someone else wants to ...
>=20
> +1
>=20
> > Another option would be to leave that for mid-beta, which is where we
> > tweaked the io_method GUCs last year. But we did that to get some
> > testing for 'worker' (in case we revert to 'sync'), and we don't need
> > that for jit.
>=20
> Doesn't seem like something to change mid-beta.  If it makes anyone
> unhappy, we'd best find out sooner not later.

I've not seen any feedback on my "counter"-proposal: switch jit_tuple_defor=
ming to off by default. Sure, for the perfect llvmjit use cases this will r=
educe the performance benefits, but it will remove most if not all the prob=
lematic queries (for instance queries running on many partitions, adding/mo=
ving columns leading to explosions in compilation time...)
Of course if there are other troublesome situations, I would love being pro=
ven wrong.