From owner-freebsd-pkgbase@freebsd.org Fri Aug 11 13:54:22 2017 Return-Path: Delivered-To: freebsd-pkgbase@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id AEF2DDD04A7 for ; Fri, 11 Aug 2017 13:54:22 +0000 (UTC) (envelope-from gjb@FreeBSD.org) Received: from freefall.freebsd.org (freefall.freebsd.org [96.47.72.132]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client CN "freefall.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 88A4672E3E; Fri, 11 Aug 2017 13:54:22 +0000 (UTC) (envelope-from gjb@FreeBSD.org) Received: from FreeBSD.org (freefall.freebsd.org [IPv6:2610:1c1:1:6074::16:84]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by freefall.freebsd.org (Postfix) with ESMTPS id 9E20213457; Fri, 11 Aug 2017 13:54:21 +0000 (UTC) (envelope-from gjb@FreeBSD.org) Date: Fri, 11 Aug 2017 13:54:19 +0000 From: Glen Barber To: Ben Woods Cc: Matthew Seaman , freebsd-pkgbase@freebsd.org Subject: Re: pkgbase update and progress Message-ID: <20170811135419.GD82820@FreeBSD.org> References: <20170810140615.GD63224@FreeBSD.org> <1218488026.1377609.1502378299872.JavaMail.zimbra@ixsystems.com> <20170810152151.GB82820@FreeBSD.org> MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; boundary="ieNMXl1Fr3cevapt" Content-Disposition: inline In-Reply-To: X-Operating-System: FreeBSD 11.0-STABLE amd64 X-SCUD-Definition: Sudden Completely Unexpected Dataloss X-SULE-Definition: Sudden Unexpected Learning Event X-PEKBAC-Definition: Problem Exists, Keyboard Between Admin/Computer X-Spidey-Sense: Uh oh, Peter logged in User-Agent: Mutt/1.8.2 (2017-04-18) X-BeenThere: freebsd-pkgbase@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: "Packaging the FreeBSD base system." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 11 Aug 2017 13:54:22 -0000 --ieNMXl1Fr3cevapt Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Aug 10, 2017 at 11:05:06PM +0000, Ben Woods wrote: > On Thu, 10 Aug 2017 at 11:35 pm, Matthew Seaman wro= te: >=20 > > On 10/08/2017 16:21, Glen Barber wrote: > > >> * Since we have no mechanism of properly merging changes to /etc > > >> we simply copy files as needed with pc-updatemanager, and it is > > >> difficult to maintain what should be updated, merged, or removed. > > >> > > > This is another long-standing item in the list of things to fix, inde= ed. > > > > The idea floated at BSDCam was to have a package that would install a > > pristine copy of the /etc files under /var/db/etcupdate and prompt the > > user to then use that tool to update or install their configs. That > > should provide a nice compromise between having the files available from > > pkgs and not enforcing a specific mechanism (tools, not policy) to > > update /etc. > > > I haven't attended either of the conferences, and I'm sure this would have > been discussed, but for those of us who weren't there: >=20 > Pkgng has 3-way merge capability (copied from fossil-scm in pkg 1.4.0). W= hy > are we considering using it? Not having ever seen this functionality used, > I would have thought the only things to do would be agree on where the > pristine old copy would be stored (.sample files in /etc?) and perhaps > improve pkgng if the user experience isn't satisfactory. >=20 The problem isn't with pkg(8) ability to do a 3-way merge, but the way the relevant files are handled and installed by the installworld and distribute targets. Last I messed with etc/Makefile, for example, everyone using a packaged base that upgraded during a 12-hour timeframe (give or take) ended up with their /etc/master.passwd being overridden. This is the kind of fallout I do not want to experience again. I have been wanting to refactor etc/Makefile for some time, but I am not confident that rewriting it to meet packaged base needs is sufficient. I think we should instead use the tools that exist to do what they are written to do, regardless of an etc/Makefile rewrite. Hopefully this clarifies it a bit more. Glen --ieNMXl1Fr3cevapt Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iQIzBAEBCAAdFiEEjRJAPC5sqwhs9k2jAxRYpUeP4pMFAlmNtwYACgkQAxRYpUeP 4pO5gA//U6UKGz80Q9n/SfPJBPyu60mx1nP6GYTRyrPi4+V9Ca8O1jMk6UC4jaE+ +0hiA4DFhnzFHWwKptlujv6fN1JacCX9LwlQY4CGl9lteALNURZXgRTlvdtAdxIK KL0+CJeOPyYfno1rxIi1tcuNvktzk1LtdstKq7q7lwhd4lyFxPH77CAL0hcIbdAI 7oIMfR1H4gES2PJ0M3Crh1UCcdsiVOEAGqx5ITW9a5GGbb0QKZV1XMJty+rnfwSQ w45E2yDosO5SmfwdlAn0s4lIBjaKhZceNJOREj0zZso+WBKWdRW1d0lEVQ2m1hbY 8zlDL4WBdRLXK+3P3nvc8UbFaxYNCjd2hSX/nfQJnx9J3oFYnJQ+ed+e9pzJpuwo cEoc6Lt3/E9OvtdSP7ncHrDfYPtmpLb0ascImTCfNd1mgDaU+jUXUgYDdipyKVBu P0wwak+07QcCAddczYNPTOpPHffTjGUn8gqzgamzgctvHo/LUXRIM0LrSV3eZsLQ szPHyZXvAV/aWvMA8Tmfu6BxzYqWoTepVhmtcULE6ywcirKUFDEDzxrc6VCIfCn3 Vvu6F5ZSjrpMiUW8an2pq944XKEWmULXcvbK46IGi6kegCglN4iP4JaiiokmpvSm S4R3k7ylbB5c1cT9MOodO/0JWQw3XAmY+S/9H8fie0Bar0H2OM4= =ldUY -----END PGP SIGNATURE----- --ieNMXl1Fr3cevapt--