From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ludovic =?UTF-8?Q?Court=C3=A8s?= Subject: bug#37506: [core-updates] Inferiors cannot reference master branch Date: Wed, 02 Oct 2019 11:40:58 +0200 Message-ID: <875zl7jxhx.fsf@gnu.org> References: <87impalx6u.fsf@gnu.org> <87muetaau0.fsf@devup.no> <87blv26vym.fsf_-_@gnu.org> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Return-path: Received: from eggs.gnu.org ([2001:470:142:3::10]:47880) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1iFb91-00036v-Qx for bug-guix@gnu.org; Wed, 02 Oct 2019 05:42:04 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1iFb90-0001Lc-Nf for bug-guix@gnu.org; Wed, 02 Oct 2019 05:42:03 -0400 Received: from debbugs.gnu.org ([209.51.188.43]:54827) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1iFb90-0001LW-Km for bug-guix@gnu.org; Wed, 02 Oct 2019 05:42:02 -0400 Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1iFb90-0008T6-Is for bug-guix@gnu.org; Wed, 02 Oct 2019 05:42:02 -0400 Sender: "Debbugs-submit" Resent-To: bug-guix@gnu.org Resent-Message-ID: In-Reply-To: <87blv26vym.fsf_-_@gnu.org> ("Ludovic \=\?utf-8\?Q\?Court\=C3\=A8s\?\= \=\?utf-8\?Q\?\=22's\?\= message of "Mon, 30 Sep 2019 10:15:45 +0200") List-Id: Bug reports for GNU Guix List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: bug-guix-bounces+gcggb-bug-guix=m.gmane.org@gnu.org Sender: "bug-Guix" To: Marius Bakke Cc: 37506-done@debbugs.gnu.org Hi, Ludovic Court=C3=A8s skribis: >> Oh I see. This is fixed by cherry-picking >> 96783ed6275cd2818ff56916274e6e4582f1dc9b from =E2=80=98core-updates=E2= =80=99 into >> =E2=80=98master=E2=80=99. Will push shortly. > > On second thought, that doesn=E2=80=99t solve the problem for revisions o= n the > current =E2=80=98master=E2=80=99 branch between and > 456c7ade095cce96f47389c37a4fe3b0390047fd and > 7089f98ef1c274f1607ec314f3a16bd3c3ac89a4. For any commit in that range, > =E2=80=98guix pull --commit=3DCOMMIT=E2=80=99 will now always fail with: [...] > ERROR: In procedure scm-error: > re-exporting local variable: AT_SYMLINK_NOFOLLOW > > It=E2=80=99s a case where code isn=E2=80=99t =E2=80=9Cforward-compatible= =E2=80=9D, and it=E2=80=99s code needed > to build the =E2=80=9Ccompute-guix-derivation=E2=80=9D trampoline. Bumme= r. > > We could hack something to work around it. For example, we could add to > =E2=80=98compiled-modules=E2=80=99 a magic parameter that would tell it t= o undefine > AT_SYMLINK_NOFOLLOW & co., or we could build the trampoline with 2.2.4. I added a =E2=80=9Cquirk=E2=80=9D for that in 37c0d4580e464a5915ea34109f788= 98325aae2c4. I suppose we=E2=80=99ll accumulate quirks like this over time here. Normal= ly they should be quite rare because the code used to build the trampoline targets the common denominator of all Guile and Guix revisions, but we=E2=80=99ll see. Ludo=E2=80=99.