Received: from out.migadu.com (out.migadu.com [91.121.223.63]) by nld3-dev1.alpinelinux.org (Postfix) with ESMTPS id 52795781AC4 for <~alpine/devel@lists.alpinelinux.org>; Tue, 3 Mar 2020 11:34:42 +0000 (UTC) Received: (Migadu outbound); Tue, 03 Mar 2020 11:34:41 +0000 Authentication-Results: out.migadu.com; auth=pass (login) Received: from [10.225.48.178] (mobile-166-170-22-112.mycingular.net [166.170.22.112]) by out.migadu.com (Haraka/2.8.16) with ESMTPSA id 271D1856-3EB5-4BBD-8ED0-6283111DBC70.1 envelope-from (authenticated bits=0) (version=TLSv1/SSLv3 cipher=ECDHE-RSA-AES256-GCM-SHA384 verify=FAIL); Tue, 03 Mar 2020 11:34:41 +0000 Date: Tue, 3 Mar 2020 06:34:37 -0500 From: Ariadne Conill To: Rasmus Thomsen , =?utf-8?Q?=7Ealpine/devel?= <~alpine/devel@lists.alpinelinux.org> Message-ID: <92c72a28-d7a1-4a50-b5cb-30e4b662411c@edison> Subject: Re: Should we install to /opt ? MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="EdoMail5e5e40cd_58a7bc09_57ba" DKIM-Signature: v=1;a=rsa-sha256;bh=3mp1LPdP2piBkC/KTjJ5Bfejq+u69Hfd/YjecCRLee0=;c=relaxed/simple;d=dereferenced.org;h=from:subject:date:to;s=default;b=S8nweqi4d92uDbtNKmEZo2RoGA8hNIQMqsxX49f0wVxBgdc59r9K2pHMnqW9mQZuFpuURbURJmFxpr9N/MVSgYU8GSGcvj+ZAdFwplB6wp3comRZHsBlY4xhFnPLklBb4iRpOpufZvJ/ha5T/LK6/cnooYtvie616vz+uWsDwyw= --EdoMail5e5e40cd_58a7bc09_57ba Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Content-Disposition: inline Hello, On Mar 3, 2020 at 5:31 AM, Rasmus Thomsen wrote: > > Hello, > > a MR for couchdb[0] was recently created that installs to /opt/alpine since couchdb puts both architecture dependent and independent files into that directory. As of right now we usually just install to /us/lib/ (e.g. for firefox and chromium), even if the package also puts architecture independent data into that directory. Should we adhere to what we've been previously been doing, or should we start adopting what FHS dictates here too and change things over to /opt/alpine/? At present, installing into /opt is flagged by abuild as a policy violation. The real question is whether /opt/$distribution should be exempted from the !fhs option. That seems reasonable to me. However before accepting CouchDB into Alpine we should create a better policy for OTP releases. For example, we do not wish to include the BEAM VM into the release, instead preferring the system package for it. Ariadne --EdoMail5e5e40cd_58a7bc09_57ba Content-Type: text/html; charset="utf-8" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline
=
Hello,

On Mar 3, 2020 at= 5:31 AM, Rasmus Thomsen <oss=40cogitri.dev> wrote: 
Hello,
a MR for couchdb=5B0=5D was recently crea=
ted that installs to /opt/alpine
since couchdb puts both architecture dependent and independent files
into that directory. As of right now we usually just install to
/us/lib/ (e.g. for firefox and chromium), even if =
the
package also puts architecture independent data into that directory.
Should we adhere to what we've been previously been doing, or should we
start adopting what =46HS dictates here too and change things over to
/opt/alpine/=3F

At present, installing into /opt is flagged = by abuild as a policy violation.  The real question is whether /opt/= =24distribution should be exempted from the =21fhs option.  That see= ms reasonable to me.

However before accepting CouchDB into Alpine we should= create a better policy for OTP releases.  =46or example, we do not = wish to include the BEAM VM into the release, instead preferring the syst= em package for it.

Ariadne
--EdoMail5e5e40cd_58a7bc09_57ba--