From owner-freebsd-ports@freebsd.org Sat Dec 2 23:20:09 2017 Return-Path: Delivered-To: freebsd-ports@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 BA88CE6C33C for ; Sat, 2 Dec 2017 23:20:09 +0000 (UTC) (envelope-from adamw@adamw.org) Received: from apnoea.adamw.org (apnoea.adamw.org [104.225.5.94]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client CN "apnoea.adamw.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 9C2D280BB3 for ; Sat, 2 Dec 2017 23:20:08 +0000 (UTC) (envelope-from adamw@adamw.org) Received: by apnoea.adamw.org (OpenSMTPD) with ESMTPSA id 85a35e85 TLS version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NO; Sat, 2 Dec 2017 16:20:00 -0700 (MST) Content-Type: text/plain; charset=us-ascii Mime-Version: 1.0 (Mac OS X Mail 11.1 \(3445.4.7\)) Subject: Re: poudriere, python ports, and flavors oh my? From: Adam Weinberger In-Reply-To: <57e06cb3-d4f5-3a7f-062a-619f536d8dc1@chrullrich.net> Date: Sat, 2 Dec 2017 16:19:58 -0700 Cc: freebsd-ports@freebsd.org Content-Transfer-Encoding: quoted-printable Message-Id: <85CBAED9-E816-461E-A409-43C4EC335BAC@adamw.org> References: <20171201175750.GA32973@vash.rhavenn.local> <57e06cb3-d4f5-3a7f-062a-619f536d8dc1@chrullrich.net> To: Christian Ullrich X-Mailer: Apple Mail (2.3445.4.7) X-BeenThere: freebsd-ports@freebsd.org X-Mailman-Version: 2.1.25 Precedence: list List-Id: Porting software to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Dec 2017 23:20:09 -0000 > On 2 Dec, 2017, at 5:55, Christian Ullrich = wrote: >=20 > * Henrik Hudson wrote: >=20 >> So, I must be missing something. I have a poudriere jail specific >> make.conf like this: >=20 >> #Python >> DEFAULT_VERSIONS+=3D python=3D3.6 python2=3D2.7 python3=3D3.6 >=20 > Same here, and this happens: >=20 > # cat /usr/local/etc/poudriere.d/py3-make.conf > DEFAULT_VERSIONS+=3D python=3D3.6 > # poudriere version > 3.2.2 > # poudriere testport -j stable -z py3 databases/py-psycopg2 > [...] > [00:00:02] Building 2 packages using 2 builders > [00:00:02] Starting/Cloning builders > [00:00:02] Hit CTRL+t at any time to see build progress and stats > [00:00:02] [01] [00:00:00] Building lang/python27 | python27-2.7.14_1 > [00:00:02] [01] [00:00:00] Finished lang/python27 | python27-2.7.14_1: = Ignored: Blacklisted > [00:00:02] [01] [00:00:00] Skipping devel/py-setuptools@py27 | = py27-setuptools-36.5.0: Dependent port lang/python27 | python27-2.7.14_1 = ignored > [00:00:02] Stopping 2 builders > [00:00:03] Error: Depends failed to build > [00:00:03] Failed ports: > [00:00:03] Skipped ports: devel/py-setuptools@py27 > [00:00:03] Cleaning up > [00:00:03] Unmounting file systems >=20 > lang/python27 is blacklisted for set py3 to prevent exactly this kind = of disaster. If I remove it from the blacklist, still _only_ = py27-flavored packages are built. >=20 > This looks like the introduction of flavors disconnected the Python = ports from DEFAULT_VERSIONS. >=20 > What is the purpose behind this, and how do I get poudriere to build = py36-flavored packages now _without_ adding "@py36" to each and every = Python port in the list I feed to bulk? >=20 > There is the claim in UPDATING that "People using Poudriere 3.2+ [...] = do not have to do anything." Unless, it appears, they use any Python = version other than 2.7, since adding flavor suffixes to port lists is = not "not anything". >=20 > My current setup is to have two sets, one with default 2.7, the other = with default 3.6, resulting in two separate pkg repos, and when I need a = Python 3.6 package, I take it from that repo. It may be that this method = is now obsolete, but I would have expected this to be mentioned = somewhere. It's a bug, and it's being worked on. amdmi3 submitted = https://reviews.freebsd.org/D13326 and it's awaiting commit. If you'd = like to test the patch and confirm that it fixes the problem that'd be = helpful. Given that most people don't participate and test patches while they're = in review, you have to expect that some bugs will only surface when it = lands in HEAD. This is precisely why we have quarterly branches, and why = we recommend that most people stay on quarterly. # Adam --=20 Adam Weinberger adamw@adamw.org https://www.adamw.org