From mboxrd@z Thu Jan 1 00:00:00 1970 From: ludo@gnu.org (Ludovic =?UTF-8?Q?Court=C3=A8s?=) Subject: bug#30879: Commit bc499b113 broke guix on guile@2.0.14, improper field initialization Date: Tue, 15 May 2018 11:20:56 +0200 Message-ID: <87wow5i8av.fsf@gnu.org> References: <20180320154302.GL105827@pe06.us.cray.com> <877eq6ibp9.fsf@gnu.org> <20180321151642.GN105827@pe06.us.cray.com> <87y3ildttr.fsf@gnu.org> <20180321211403.GO105827@pe06.us.cray.com> <87r2oddo9l.fsf@gnu.org> <20180322144538.GP105827@pe06.us.cray.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Return-path: Received: from eggs.gnu.org ([2001:4830:134:3::10]:49801) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fIW9m-0006iA-ML for bug-guix@gnu.org; Tue, 15 May 2018 05:22:07 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fIW9h-0003B8-VZ for bug-guix@gnu.org; Tue, 15 May 2018 05:22:06 -0400 Received: from debbugs.gnu.org ([208.118.235.43]:55326) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1fIW9h-0003B4-Rv for bug-guix@gnu.org; Tue, 15 May 2018 05:22:01 -0400 Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1fIW9h-0000M4-Mk for bug-guix@gnu.org; Tue, 15 May 2018 05:22:01 -0400 Sender: "Debbugs-submit" Resent-Message-ID: In-Reply-To: <20180322144538.GP105827@pe06.us.cray.com> (Eric Bavier's message of "Thu, 22 Mar 2018 09:45:38 -0500") 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: Eric Bavier Cc: 30879@debbugs.gnu.org Hello Eric, Sorry for the late reply. Eric Bavier skribis: > On Thu, Mar 22, 2018 at 12:04:06AM +0100, Ludovic Court=C3=A8s wrote: >> Eric Bavier skribis: >>=20 >> [...] >>=20 >> > In gnu/system.scm: >> > 501: 3 [operating-system-services # # #f] >> > 476: 2 [essential-services # # #f] >> > 576: 1 [operating-system-etc-service #] >> > In gnu/system/nss.scm: >> > 217: 0 [name-service-switch->string (# # # # ...)] >> > >> > gnu/system/nss.scm:217:19: In procedure name-service-switch->string: >> > gnu/system/nss.scm:217:19: In procedure struct_vtable: Wrong type argu= ment in position 1 (expecting struct): (#< type: # Oh, so it loks like .go files from the system-installed guix are being > picked up: > > 53692 openat(AT_FDCWD, "/usr/local/lib/guile/2.0/site-ccache/gnu/system.g= o", O_RDONLY|O_CLOEXEC) =3D 10 > > I hadn't expected that, but I suppose it makes sense. Running make > under ./pre-inst-env does not help. > > We should probably find a way to prevent this in general, right? It seems that the problem here is that both Guile and Guix were installed with --prefix=3D/usr/local. Guile contains by default $prefix/lib/guile/2.0/site-ccache in its %load-compiled-path. Thus, it will always find the .go files of that Guix that=E2=80=99s installed in the same prefix. > We shouldn't be loading guix modules from outside the source tree > during build. In general we can (and do: see the =E2=80=98make-go=E2=80=99 target and see =E2=80=98pre-inst-env=E2=80=99), but in this case we can=E2=80=99t really p= revent it because $prefix/lib/=E2=80=A6 is in the default search path of Guile, which is admittedly problematic. Maybe we should just forbid install Guix in the same prefix as Guile, and detect that at configure time. WDYT? Ludo=E2=80=99.