MIME-Version: 1.0
References: 
 <CAAAe_zBi1dOtWb2vnwSvGwuU0-bqAOm_7dOM4u-CmukA8xaV5Q@mail.gmail.com>
 <CAAAe_zAuHwqUfqJOD4PDUkWsxTfTytNaandq11Kddw2bfCcpvQ@mail.gmail.com>
 <p7r5bekdbl2zcazid7agvfo2nfnq5bim2a5jkckqygld32n325@fctfp6ou6qnb>
 <20260605.074927.1218738044739683444.ishii@postgresql.org>
In-Reply-To: <20260605.074927.1218738044739683444.ishii@postgresql.org>
Reply-To: assam258@gmail.com
From: Henson Choi <assam258@gmail.com>
Date: Fri, 5 Jun 2026 08:24:05 +0900
Message-ID: 
 <CAAAe_zBdR_23drKa38cyejE1YZ+4PtasVXGJ++uSeqpvXgJeOQ@mail.gmail.com>
Subject: Re: Row pattern recognition
To: Tatsuo Ishii <ishii@postgresql.org>, andres@anarazel.de
Cc: jian.universality@gmail.com, zsolt.parragi@percona.com,
	sjjang112233@gmail.com, vik@postgresfriends.org, er@xs4all.nl,
	jacob.champion@enterprisedb.com, david.g.johnston@gmail.com,
	peter@eisentraut.org, li.evan.chao@gmail.com, pgsql-hackers@postgresql.org
Content-Type: multipart/alternative; boundary="000000000000d377ad065375d7c0"
Archived-At: 
 <https://www.postgresql.org/message-id/CAAAe_zBdR_23drKa38cyejE1YZ%2B4PtasVXGJ%2B%2BuSeqpvXgJeOQ%40mail.gmail.com>
Precedence: bulk

--000000000000d377ad065375d7c0
Content-Type: text/plain; charset="UTF-8"

> Henson,
> Do you have any idea?

Let me be honest about this one.

I think the root issue is that none of us working on the patch has
enough confidence -- enough real control -- over the tuple slot and JIT
side of it. I wrote the code, it ran fine for me, and I assumed that
meant it was fine; but I have never felt that I fully understand and
control that area. So I cannot point to a specific cause with confidence
yet, and I would rather say that plainly than guess.

One practical note: there have been no tuple slot or JIT changes in the
patch since v47, and none are planned. So that part has been stable, and
reviewing the tuple slot / JIT code against v47 is perfectly viable --
there is no need to wait for a newer revision to look at that area.

What I think would genuinely harden the patch is for someone who knows
tuple slots and JIT well to review -- or rewrite -- that part. I would
very much welcome that, and I am happy to provide the design direction
behind the current code, or any other information a reviewer would need.

To be fair, the same caveat applies to the existing planner code the
patch builds on. I feel I have a reasonable grip on the rest of the
patch, but not on those two areas.

I will keep digging on my side (trying to reproduce the JIT crash with
JIT forced on against the current versions), and report back with
anything concrete I find.

Regards,
Henson

--000000000000d377ad065375d7c0
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">&gt; Henson,<br>&gt; Do you have any idea?<br><br>Let me b=
e honest about this one.<br><br>I think the root issue is that none of us w=
orking on the patch has<br>enough confidence -- enough real control -- over=
 the tuple slot and JIT<br>side of it. I wrote the code, it ran fine for me=
, and I assumed that<br>meant it was fine; but I have never felt that I ful=
ly understand and<br>control that area. So I cannot point to a specific cau=
se with confidence<br>yet, and I would rather say that plainly than guess.<=
br><br>One practical note: there have been no tuple slot or JIT changes in =
the<br>patch since v47, and none are planned. So that part has been stable,=
 and<br>reviewing the tuple slot / JIT code against v47 is perfectly viable=
 --<br>there is no need to wait for a newer revision to look at that area.<=
br><br>What I think would genuinely harden the patch is for someone who kno=
ws<br>tuple slots and JIT well to review -- or rewrite -- that part. I woul=
d<br>very much welcome that, and I am happy to provide the design direction=
<br>behind the current code, or any other information a reviewer would need=
.<br><br>To be fair, the same caveat applies to the existing planner code t=
he<br>patch builds on. I feel I have a reasonable grip on the rest of the<b=
r>patch, but not on those two areas.<br><br>I will keep digging on my side =
(trying to reproduce the JIT crash with<br>JIT forced on against the curren=
t versions), and report back with<br>anything concrete I find.<br><br>Regar=
ds,<br>Henson<br></div>

--000000000000d377ad065375d7c0--