From mboxrd@z Thu Jan 1 00:00:00 1970 From: ludo@gnu.org (Ludovic =?utf-8?Q?Court=C3=A8s?=) Subject: Re: [PATCH 01/10] * gnu/system.scm (): Add 'store-device' and 'store-fs-mount-point'. Date: Sun, 30 Oct 2016 23:19:56 +0100 Message-ID: <87d1ihscmr.fsf@gnu.org> References: <20161028100727.1182-1-cmmarusich@gmail.com> <20161028100727.1182-2-cmmarusich@gmail.com> <87h97uu23b.fsf@gnu.org> <87lgx6fa24.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 ([2001:4830:134:3::10]:34964) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1c0ySQ-0001TA-QV for guix-devel@gnu.org; Sun, 30 Oct 2016 18:20:03 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1c0ySN-0004zK-KF for guix-devel@gnu.org; Sun, 30 Oct 2016 18:20:02 -0400 In-Reply-To: <87lgx6fa24.fsf@gmail.com> (Chris Marusich's message of "Sun, 30 Oct 2016 02:41:39 -0700") 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" To: Chris Marusich Cc: guix-devel@gnu.org Chris Marusich skribis: > ludo@gnu.org (Ludovic Court=C3=A8s) writes: >> However, after rereading the whole series, I think what we need is a >> =E2=80=98device-mount-point=E2=80=99 in , in a symmetrical f= ashion. > > That would work, and I'm open to it. However, before we decide to do > that, what do you think about the alternative in which we simply require > that all the paths in the and objects > must be relative paths? By relative, I mean relative to the GRUB root. > If all the paths were relative, then we would not need a > 'device-mount-point' field in the first place. > > However, I suspect it might be tricky to get all the information we need > if we do it that way. For example, during switch-generation, how would > we know the correct paths to use for the files in the GRUB > configuration's 'eye-candy' section? In light of that, I suspect it > would be better to do what you've suggested: include the mount point in > both and . Yeah, I=E2=80=99m not sure we could easily manage to have only relative file names because most of them are computed when the derivations get built (that=E2=80=99s why =E2=80=98strip-mount-point=E2=80=99 currently has to re= turn a gexp instead of stripping the mount point directly.) >> Attached is an updated patch. I have grouped together the patches of >> your series that touch this topic, including the documentation part >> (this is all one logical change so it=E2=80=99s best to commit it as suc= h), and >> made the above change. > > Thank you for doing this. I've looked it over. It looks good to me. > Unless you think strongly that we should use relative paths in the > and objects, I think it's good as-is! Good! >> If that=E2=80=99s fine with you, I=E2=80=99d like to commit this version= . With that >> done, the rest of the patch series will be rather easy. >> >> WDYT? >> > > Sounds good to me. Pushed as 1ef8b72a7f87afe7cffe52393d99e1b14e4770e1! > Unless you think strongly that we should use relative paths in the > and objects, I will re-submit my > patches once you have committed this change to the official repo. That would be awesome. Thanks again for all your work and patience! Ludo=E2=80=99.