From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?UTF-8?B?SmFuIFN5bsOhxI1law==?= Subject: Re: Test failure when building libarchive-3.1.2 Date: Sat, 21 Nov 2015 08:58:51 +0100 Message-ID: References: <87h9koyfkr.fsf@gnu.org> <87twogixth.fsf@gnu.org> <87bnaoe19d.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 ([2001:4830:134:3::10]:39149) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1a034Q-0003aw-0w for guix-devel@gnu.org; Sat, 21 Nov 2015 02:58:55 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1a034O-0005Cu-Sq for guix-devel@gnu.org; Sat, 21 Nov 2015 02:58:53 -0500 In-Reply-To: <87bnaoe19d.fsf@gnu.org> List-Id: "Development of GNU Guix and the GNU System distribution." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org Sender: guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org To: =?UTF-8?Q?Ludovic_Court=C3=A8s?= Cc: guix-devel On Fri, Nov 20, 2015 at 11:47 PM, Ludovic Court=C3=A8s wrote= : > Jan Syn=C3=A1=C4=8Dek skribis: > >> On Fri, Nov 20, 2015 at 2:51 PM, Ludovic Court=C3=A8s wro= te: >>> Jan Syn=C3=A1=C4=8Dek skribis: >>> >>>> On Sat, Nov 14, 2015 at 12:40 PM, Ludovic Court=C3=A8s = wrote: > > [...] > >>>>> It would be nice to see if this systematically fails. If it is >>>>> non-deterministic, we should build it with --keep-failed until it fai= ls >>>>> (removing successful builds with =E2=80=98guix gc -d=E2=80=99), colle= ct useful info from >>>>> the build tree, and debug. > > [...] > >>> I noticed that libarchive uses =E2=80=98readdir=E2=80=99 calls as-is, w= ithout sorting >>> directory entries afterwards. Thus, the order of directory entries is >>> effectively non-deterministic and may change depending on the phase of >>> the moon. >>> >>> This has been reported at: >>> >>> https://github.com/libarchive/libarchive/issues/602 >>> >>> Could you add the patch that=E2=80=99s given at that URL to the =E2=80= =98patches=E2=80=99 field >>> or libarchive=E2=80=99s =E2=80=98origin=E2=80=99 form and see if the pr= oblem shows up again, >>> preferably building several times in a row? >> >> I built it once and it passed (note that it failed *everytime* I >> wanted to build it). > > So this patch appears to solve the issue? > >> Maybe a dumb question, but how do I force a rebuild of an already >> built package?:) > > You can=E2=80=99t rebuild packages because build processes are assumed to= be > deterministic. However, you can delete a build result with =E2=80=98gc -= d=E2=80=99, as > mentioned above, and rebuild it afterwards. So I've built the package several times and it worked every time. I don't e= ven think the build was failing non-deterministically, because, as the commit message of the patch explains, the unpatched bsdtar was trying to put things like SELinux contexts into the archive, which resulted in an archive with bigger size th= an anticipated. And since I'm running guix on Fedora, I probably hit the SELinux problem. Thank you for helping out! Cheers, --=20 Jan Syn=C3=A1=C4=8Dek