From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ludovic =?UTF-8?Q?Court=C3=A8s?= Subject: bug#35543: Installer crashes on retry when selecting keyboard layout Date: Fri, 03 May 2019 15:41:11 +0200 Message-ID: <875zqrd5ag.fsf@gnu.org> References: <20190503090937.vveultdxob2fzliv@pelzflorian.localdomain> <87imurd9vu.fsf@gnu.org> <874l6bka82.fsf@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Return-path: Received: from eggs.gnu.org ([209.51.188.92]:35808) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1hMYRv-0008GH-Aw for bug-guix@gnu.org; Fri, 03 May 2019 09:42:04 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1hMYRu-00007Q-II for bug-guix@gnu.org; Fri, 03 May 2019 09:42:03 -0400 Received: from debbugs.gnu.org ([209.51.188.43]:34565) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1hMYRu-00007L-Dw for bug-guix@gnu.org; Fri, 03 May 2019 09:42:02 -0400 Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1hMYRu-0006s7-A2 for bug-guix@gnu.org; Fri, 03 May 2019 09:42:02 -0400 Sender: "Debbugs-submit" Resent-Message-ID: In-Reply-To: <874l6bka82.fsf@gmail.com> (Mathieu Othacehe's message of "Fri, 03 May 2019 14:12:45 +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: Mathieu Othacehe Cc: 35543@debbugs.gnu.org Hello! Mathieu Othacehe skribis: >> I believe the broader issue is that the installer cannot really be >> restarted after a failed install, as we discussed a while back on >> guix-devel. Mathieu, can you confirm? > > I'm not aware of this layout issue, however, after the installer is > restarted, the install is doomed because some mounts can't be > undone. > > I didn't find a way to fix this, so maybe we should advise/force a > reboot upon installer failure? Maybe, though we should also offer a way to collect more debugging info and report the issue, which means at least letting people read error messages. Thoughts? Ludo=E2=80=99.