From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:470:142:3::10]:33531) by lists.gnu.org with esmtp (Exim 4.86_2) (envelope-from ) id 1hZEj7-0006bE-4e for guix-patches@gnu.org; Fri, 07 Jun 2019 09:16:15 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1hZEj5-0005Ed-7p for guix-patches@gnu.org; Fri, 07 Jun 2019 09:16:13 -0400 Received: from debbugs.gnu.org ([209.51.188.43]:37993) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1hZEix-0004wc-2o for guix-patches@gnu.org; Fri, 07 Jun 2019 09:16:05 -0400 Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1hZEiw-0003rI-BM for guix-patches@gnu.org; Fri, 07 Jun 2019 09:16:02 -0400 Subject: [bug#36093] [PATCH 2/2] pack: Add '--entry-point'. Resent-Message-ID: From: Ludovic =?UTF-8?Q?Court=C3=A8s?= References: <20190604210115.24477-1-ludo@gnu.org> <20190604210115.24477-2-ludo@gnu.org> <20190605170653.5be950ca@scratchpost.org> <87blzbydzk.fsf@gnu.org> <87ftolk85v.fsf@mdc-berlin.de> Date: Fri, 07 Jun 2019 15:15:39 +0200 In-Reply-To: <87ftolk85v.fsf@mdc-berlin.de> (Ricardo Wurmus's message of "Fri, 7 Jun 2019 12:21:32 +0200") Message-ID: <87imthedtw.fsf@gnu.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: guix-patches-bounces+kyle=kyleam.com@gnu.org Sender: "Guix-patches" To: Ricardo Wurmus Cc: 36093@debbugs.gnu.org Hello, Ricardo Wurmus skribis: > Ludovic Court=C3=A8s writes: > >>> Also, in general, do we conflate "squashfs" and "singularity"? It has = been >>> that way in guix/scripts/pack.scm's squashfs-image before this patch al= ready >>> and a few extra files can't hurt, but we could also just provide a >>> function "singularity-image" or something. >> >> Yes, we do conflate Singularity and Squashfs, but I think there=E2=80=99= s no >> other =E2=80=9Ccontainer tool=E2=80=9D that uses Squashfs anyway. > > When I originally added the squashfs support to =E2=80=9Cguix pack=E2=80= =9D I had > Singularity in mind, but since it didn=E2=80=99t do anything particular f= or > Singularity I named it =E2=80=9Csquashfs=E2=80=9D. > > squashfs is used as a format by Snap (which we don=E2=80=99t explicitly s= upport > yet), but it is also generally useful as a way to share disk images, > which could for example be used with lxc containers. Oh, I didn=E2=80=99t know LXC and Snap support squashfs. >> We could rename it to =E2=80=9Csingularity=E2=80=9D, but it turns out Si= ngularity 3.x >> has its own image format unimaginatively called SIF, so perhaps we=E2=80= =99re >> better off with the status quo. >> >> Thoughts? Ricardo? > > In my opinion, going forward we should not conflate =E2=80=9Csquashfs=E2= =80=9D and > Singularity more and eventually *add* a format handler for Singularity > 3.x. > > But these changes to the =E2=80=9Csquashfs=E2=80=9D format handler look f= ine to me. > Let=E2=80=99s deal with Singularity 3.x later. What about: 1. Renaming =E2=80=98squashfs=E2=80=99 to =E2=80=98singularity-squashfs= =E2=80=99, and deprecating =E2=80=98squashfs=E2=80=99. 2. Eventually, add a =E2=80=98sif=E2=80=99 format for Singularity 3=E2=80= =99s native image format. 3. Add a =E2=80=98snap=E2=80=99 backend, and perhaps an =E2=80=98lxc=E2= =80=99 backend too. Thanks, Ludo=E2=80=99.