From debbugs-submit-bounces@debbugs.gnu.org Wed Feb 19 09:21:08 2020 Received: (at 39266) by debbugs.gnu.org; 19 Feb 2020 14:21:08 +0000 Received: from localhost ([127.0.0.1]:41835 helo=debbugs.gnu.org) by debbugs.gnu.org with esmtp (Exim 4.84_2) (envelope-from ) id 1j4QDs-0002Fb-9k for submit@debbugs.gnu.org; Wed, 19 Feb 2020 09:21:08 -0500 Received: from mail.inskydata.com ([184.68.229.6]:54948 helo=server.inskydata.com) by debbugs.gnu.org with esmtp (Exim 4.84_2) (envelope-from ) id 1j4QDo-0002EH-IF for 39266@debbugs.gnu.org; Wed, 19 Feb 2020 09:21:06 -0500 Received: from [192.168.34.129] (unknown [192.168.34.129]) by server.inskydata.com (Postfix) with ESMTPSA id 2F1335741D0F; Wed, 19 Feb 2020 07:20:34 -0700 (MST) Content-Type: text/plain; charset=utf-8 Mime-Version: 1.0 (1.0) Subject: Re: bug#39266: Finalization thread hits wrong-type-arg on weak vector (AArch64) From: Brian Woodcox X-Mailer: iPhone Mail (16F203) In-Reply-To: <87k14in0sk.fsf@gnu.org> Date: Wed, 19 Feb 2020 07:19:57 -0700 Content-Transfer-Encoding: quoted-printable Message-Id: <90401426-AF73-418E-89DD-1A22FC704BD7@InSkyData.com> References: <87tv4kdgyy.fsf@inria.fr> <87k14in0sk.fsf@gnu.org> To: =?utf-8?Q?Ludovic_Court=C3=A8s?= X-Spam-Score: 0.3 (/) X-Debbugs-Envelope-To: 39266 Cc: 39266@debbugs.gnu.org X-BeenThere: debbugs-submit@debbugs.gnu.org X-Mailman-Version: 2.1.18 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: debbugs-submit-bounces@debbugs.gnu.org Sender: "Debbugs-submit" X-Spam-Score: -1.0 (-) Well, I would be happy, because right now I can=E2=80=99t do guix pull. I tried multiple timed on aarch64 with no success. Brian. > On Feb 19, 2020, at 6:50 AM, Ludovic Court=C3=A8s wrote: >=20 > Ludovic Court=C3=A8s skribis: >=20 >> While building the =E2=80=9Cguix-system.drv=E2=80=9D derivation on AArch6= 4, I got this >> crash (not fully deterministic but quite frequent). Here the >> finalization thread gets a wrong-type-arg in =E2=80=98scm_i_weak_car=E2=80= =99 (i.e., >> accessing a one-element weak vector): >>=20 >> $ ( export out=3D$PWD/build; unset GUILE_LOAD_PATH; unset GUILE_LOAD_COMP= ILED_PATH; gdb --args "/gnu/store/p8in2npgl5yhliy25ikz7shjbq0gii95-guile-nex= t-3.0.0/bin/guile" "--no-auto-compile" "-L" "/gnu/store/3qg8l6kr4wa9sbgwy00z= 1mb3p88xf455-module-import" "-C" "/gnu/store/h9qcvg71bmx735fsndagll9y7s72k9n= 9-module-import-compiled" guix-system-builder ) >> [=E2=80=A6] >> loading 'gnu/services/cups.scm'... >> Backtrace: >=20 > Apparently this bug does not occur with v3.0.0-23-g7dc90a17e=C2=B9. It ma= y > be that 00fbdfa7345765168e14438eed0b0b8c64c27ab9 reduces GC pressure, > which as a side effect makes the problem vanish. >=20 > It=E2=80=99s not satisfactory, but as a stop-gap measure, we could release= 3.0.1 > like this, which could make Guile 3 usable for Guix on AArch64. >=20 > Thoughts? >=20 > Ludo=E2=80=99. >=20 > =C2=B9 Specifically, I tested by (1) building a tarball with =E2=80=9Cmake= dist=E2=80=9D, (2) > running =E2=80=9Cguix build guile-next > --with-source=3Dguile-next=3Dthe-tarball.tar.gz=E2=80=9D, and (3) running= that > Guile in the code above. For some reason, Guile 3.0.0 built =E2=80=9Cby h= and=E2=80=9D > would not reproduce the original bug, which is why I built it through > Guix. >=20 >=20 >=20