MIME-Version: 1.0
References: <CADfhSr-pCkbDxmiOVYSAGE5QGjsQ48KKH_W424SPk+pwzKZFaQ@mail.gmail.com>
 <CAFY6G8cgNNw86oPAE6MUWm=sPd7T0xr+5qBJWEtvre2n-wUetg@mail.gmail.com>
 <CAEG8a3JR8quniKWNu4jTJJM_1fNiCPnsEQPCDO+NFtKup7MTbQ@mail.gmail.com>
 <CA+HiwqF+pBaQ5tnH7Mu=7rsNUsq1LSD9WvzCeQJz=2qD5unB0A@mail.gmail.com> <CA+HiwqG_PXQqL9TvQFEdyJfB9Wq1-Wm3SFppWHhMg+z2gD_nag@mail.gmail.com>
In-Reply-To: <CA+HiwqG_PXQqL9TvQFEdyJfB9Wq1-Wm3SFppWHhMg+z2gD_nag@mail.gmail.com>
From: Fredrik Widlert <fredrik.widlert@digpro.se>
Date: Fri, 10 Apr 2026 09:29:44 +0200
Message-ID: <CADfhSr8hYc-4Cz7vfXH_oV-Jq81pyK9W4phLrOGspovsg2W7Kw@mail.gmail.com>
Subject: Re: BUG: PostgreSQL 19devel throws internal opfamily error for FK
 with reordered referenced columns
To: Amit Langote <amitlangote09@gmail.com>
Cc: Junwang Zhao <zhjwpku@gmail.com>, Matheus Alcantara <matheusssilv97@gmail.com>, 
	pgsql-bugs@lists.postgresql.org
Content-Type: multipart/alternative; boundary="000000000000814538064f161980"
Archived-At: <https://www.postgresql.org/message-id/CADfhSr8hYc-4Cz7vfXH_oV-Jq81pyK9W4phLrOGspovsg2W7Kw%40mail.gmail.com>
Precedence: bulk

--000000000000814538064f161980
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Thanks for the very quick fix - it's a great feeling to report a problem
and get it fixed this quickly.

I can confirm that the fix solves not only the small reproducer but also
the original problems I got when trying to install our (large) system
on PostgreSQL 19 devel for testing.

We actually ran into the same problem in two ways:
* the reordered FK with datatype mismatch (my reproducer)
* other reordered FKs with the same datatypes, where we instead got
  regular FK violations

In case you want a reproducer for the second case as well, I've included it
below,
but perhaps it's not relevant since the patch already fixed it.

----
drop table if exists parent, child;

create table parent (
     c1 integer,
     c2 integer,
     primary key (c1, c2)
);

create table child (
     c1 integer,
     c2 integer,
     constraint child_fk foreign key (c2, c1) references parent (c2, c1)
);

insert into parent (c1, c2) values (1, 2);
insert into child (c1, c2) values (1, 2);

/Fredrik Widlert

On Fri, Apr 10, 2026 at 8:32=E2=80=AFAM Amit Langote <amitlangote09@gmail.c=
om>
wrote:

> On Fri, Apr 10, 2026 at 1:42=E2=80=AFPM Amit Langote <amitlangote09@gmail=
.com>
> wrote:
> > Thanks Junwang for checking.  I've just pushed this: 980c1a85d819.
>
> Pushed a follow up commit to add an Assert making the single-column
> invariant explicit in ri_FastPathFlushArray(), per Junwang's
> suggestion off-list.
>
> --
> Thanks, Amit Langote
>

--000000000000814538064f161980
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Thanks for the very quick fix - it&#39;s a great feeling t=
o report a problem<br>and get it fixed this quickly.<br><br>I can confirm t=
hat the fix solves not only the small reproducer but also<br>the original p=
roblems I got when trying to install our (large) system<br>on PostgreSQL 19=
 devel for testing.<br><br>We actually ran into the same problem in two way=
s:<br>* the reordered FK with datatype mismatch (my reproducer)<br>* other =
reordered FKs with the same datatypes, where we instead got<br>=C2=A0 regul=
ar FK violations<br><br>In case you want a reproducer for the second case a=
s well, I&#39;ve included it below,<br>but perhaps it&#39;s not relevant si=
nce the patch already fixed it.<br><br>----<br>drop table if exists parent,=
 child;<br><br>create table parent (<br>=C2=A0 =C2=A0 =C2=A0c1 integer,<br>=
=C2=A0 =C2=A0 =C2=A0c2 integer,<br>=C2=A0 =C2=A0 =C2=A0primary key (c1, c2)=
<br>);<br><br>create table child (<br>=C2=A0 =C2=A0 =C2=A0c1 integer,<br>=
=C2=A0 =C2=A0 =C2=A0c2 integer,<br>=C2=A0 =C2=A0 =C2=A0constraint child_fk =
foreign key (c2, c1) references parent (c2, c1)<br>);<br><br>insert into pa=
rent (c1, c2) values (1, 2);<br>insert into child (c1, c2) values (1, 2);<b=
r><br>/Fredrik Widlert</div><br><div class=3D"gmail_quote gmail_quote_conta=
iner"><div dir=3D"ltr" class=3D"gmail_attr">On Fri, Apr 10, 2026 at 8:32=E2=
=80=AFAM Amit Langote &lt;<a href=3D"mailto:amitlangote09@gmail.com">amitla=
ngote09@gmail.com</a>&gt; wrote:<br></div><blockquote class=3D"gmail_quote"=
 style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);p=
adding-left:1ex">On Fri, Apr 10, 2026 at 1:42=E2=80=AFPM Amit Langote &lt;<=
a href=3D"mailto:amitlangote09@gmail.com" target=3D"_blank">amitlangote09@g=
mail.com</a>&gt; wrote:<br>
&gt; Thanks Junwang for checking.=C2=A0 I&#39;ve just pushed this: 980c1a85=
d819.<br>
<br>
Pushed a follow up commit to add an Assert making the single-column<br>
invariant explicit in ri_FastPathFlushArray(), per Junwang&#39;s<br>
suggestion off-list.<br>
<br>
-- <br>
Thanks, Amit Langote<br>
</blockquote></div>

--000000000000814538064f161980--