MIME-Version: 1.0
References: <E1w0jP7-003dld-0s@gemulon.postgresql.org>
 <CAApHDvquH6wnp4fhpaCOkC4R3KAvr2BOTbhhDPDQCBNR3YbLMQ@mail.gmail.com>
In-Reply-To: 
 <CAApHDvquH6wnp4fhpaCOkC4R3KAvr2BOTbhhDPDQCBNR3YbLMQ@mail.gmail.com>
From: Robert Haas <robertmhaas@gmail.com>
Date: Tue, 17 Mar 2026 09:59:10 -0400
Message-ID: 
 <CA+TgmoZw5tx_Jyxr-KeNAXsPV1yQ4ducgieUxY6+ghd33tCn9Q@mail.gmail.com>
Subject: Re: pgsql: Add pg_plan_advice contrib module.
To: David Rowley <dgrowleyml@gmail.com>
Cc: 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%2BTgmoZw5tx_Jyxr-KeNAXsPV1yQ4ducgieUxY6%2Bghd33tCn9Q%40mail.gmail.com>
Precedence: bulk

On Mon, Mar 16, 2026 at 5:12=E2=80=AFPM David Rowley <dgrowleyml@gmail.com>=
 wrote:
> I was just building with VS2026 when I noticed the following warning:
>
> ../contrib/pg_plan_advice/pgpa_planner.c(1195): warning C4319: '~':
> zero extending 'uint32' to 'uint64' of greater size
>
> I guess the compiler is concerned that you might not have meant to
> unconditionally switch off all the upper 32-bits of the 64-bit
> variable.  Adding the following does get rid of the warning for me,
> but not sure if that's what you had intended. I've not studied the
> code enough to understand why there are two different integer widths
> being used for the bits.
>
> -               *pgs_mask_p &=3D ~(PGS_JOIN_ANY & ~join_mask);
> +               *pgs_mask_p &=3D (uint32) ~(PGS_JOIN_ANY & ~join_mask);

Thanks for the report. This is an error on my part, pure and simple.
At one point, pgs_mask was a 32-bit value, and although that changed
months ago, this holdover escaped detection. I'll go fix that.

--=20
Robert Haas
EDB: http://www.enterprisedb.com