MIME-Version: 1.0
References: <CA+hUKGJTumad75o8Zao-LFseEbt=enbUFCM7LZVV=c8yg2i7dg@mail.gmail.com>
 <CA+hUKGJv+QUOy8yOHe_0VfRGTSrHaCvo0xZG-1JGTqBurcPkOw@mail.gmail.com> <CA+hUKGLJ63fpt_u0OXDR91fbpicDXK01fivzDSBMOtPazEanFQ@mail.gmail.com>
In-Reply-To: <CA+hUKGLJ63fpt_u0OXDR91fbpicDXK01fivzDSBMOtPazEanFQ@mail.gmail.com>
From: Thomas Munro <thomas.munro@gmail.com>
Date: Thu, 22 Jan 2026 15:24:56 +1300
Message-ID: <CA+hUKGJNP49kDT0WZph3MwBdYzwy7WnYCpy+XNJpsO78AxXydw@mail.gmail.com>
Subject: Re: LLVM 22
To: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Cc: Matheus Alcantara <matheusssilv97@gmail.com>, 
	Anthonin Bonnefoy <anthonin.bonnefoy@datadoghq.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: <https://www.postgresql.org/message-id/CA%2BhUKGJNP49kDT0WZph3MwBdYzwy7WnYCpy%2BXNJpsO78AxXydw%40mail.gmail.com>
Precedence: bulk

On Sun, Jan 11, 2026 at 8:09=E2=80=AFPM Thomas Munro <thomas.munro@gmail.co=
m> wrote:
> A new unrelated assertion started firing in LLVM main/22 a few days ago:
>
>     v_nullbytemask =3D l_int8_const(lc, 1 << ((attnum) & 0x07));
>     Assertion failed: (llvm::isUIntN(BitWidth, val) && "Value is not
> an N-bit unsigned value")
>
> Here is a fix for that.

22 was branched and RC1 is out, but that particular change was
reverted from 22[1].  It had already been through a commit/revert
cycle before and at a wild guess, it probably caused too much work
elsewhere with not enough notice.  It's still present in main, so
consider the v2-0003 patch booted out of here and into the
not-yet-created LLVM 23 thread...

[1] https://github.com/llvm/llvm-project/commit/16bf1c5d6b7f8fda16da5df5a2b=
195a6b10d08ed