MIME-Version: 1.0
References: 
 <CAHg+QDcd=t69gLf9yQexO07EJ2mx0Z70NFHo6h94X1EDA=hM0g@mail.gmail.com>
 <CACJufxGreOtA-S-qeHyS5iSSsj5zZX0W3Rf8FxbyL+SVXFjLYw@mail.gmail.com>
 <CAHg+QDeGLfz8YSCChjqrxaVSrz9AnMA0NrmsNogLqeGgCt7-wg@mail.gmail.com>
 <CA+renyWD+XXifwswE74vhjooqbiVKu4qVhLvpMcUQBzrjVjT7A@mail.gmail.com>
 <CACJufxHYntqy2fo9CFWDDrqKjcMK8DGRM3kse4YnXYnPYq2Hiw@mail.gmail.com>
 <CA+renyVp4rgj8x0ERXRkZp223eyBZ_XZr2RVCXvjzKBhTtS6Yw@mail.gmail.com>
 <CACJufxEkomKYmWgqXJmQr_qS+z=BZ3w801eh7Z7ekh-3oHXxHQ@mail.gmail.com>
In-Reply-To: 
 <CACJufxEkomKYmWgqXJmQr_qS+z=BZ3w801eh7Z7ekh-3oHXxHQ@mail.gmail.com>
From: Paul A Jungwirth <pj@illuminatedcomputing.com>
Date: Tue, 21 Apr 2026 07:59:56 -0700
Message-ID: 
 <CA+renyWk7kVsZJPZKzN95mYkO7S=hDUx=+fUPtbg9qFqeepCpg@mail.gmail.com>
Subject: Re: FOR PORTION OF does not recompute GENERATED STORED columns that
 depend on the range column
To: jian he <jian.universality@gmail.com>
Cc: SATYANARAYANA NARLAPURAM <satyanarlapuram@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/CA%2BrenyWk7kVsZJPZKzN95mYkO7S%3DhDUx%3D%2BfUPtbg9qFqeepCpg%40mail.gmail.com>
Precedence: bulk

On Mon, Apr 20, 2026 at 8:58=E2=80=AFPM jian he <jian.universality@gmail.co=
m> wrote:
>
> + updatedCols =3D
> + bms_add_member(updatedCols,
> +   rangeAttno - FirstLowInvalidHeapAttributeNumber);
>
> Here, use "perminfo->updatedCols" not "updatedCols", otherwise segfault.
> The attached diff based on v5, fixes this issue.
>
> +ALTER TABLE temporal_partitioned_3 ADD COLUMN range_len int GENERATED
> ALWAYS AS (upper(valid_at) - lower(valid_at)) STORED;
> Slightly refactoring the tests will allow for easier comparison of
> range_len values.

I can't reproduce a segfault. I don't see how that code would lead to
one. Can you share what you're doing to cause it? Simply running the
regression tests doesn't do it for me.

The v5 patch has quite a lot of code shared between both branches. I
was trying to avoid that in the v5 patch. Is there any way to clean it
up?

I like the how the test change makes it easy to verify the range_len column=
.

Yours,

--=20
Paul              ~{:-)
pj@illuminatedcomputing.com