From owner-svn-ports-head@freebsd.org Thu Apr 27 17:09:51 2017 Return-Path: Delivered-To: svn-ports-head@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 DAC8BD53457; Thu, 27 Apr 2017 17:09:51 +0000 (UTC) (envelope-from vsevolod@FreeBSD.org) Received: from mail.highsecure.ru (mail6.highsecure.ru [IPv6:2a01:4f8:190:43b5::99]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id A0530AA9; Thu, 27 Apr 2017 17:09:51 +0000 (UTC) (envelope-from vsevolod@FreeBSD.org) Received: from thinkpad.local (unknown [81.145.134.164]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) (Authenticated sender: thinkpad@highsecure.ru) by mail.highsecure.ru (Postfix) with ESMTPSA id 826B730013F; Thu, 27 Apr 2017 19:09:39 +0200 (CEST) Received: from [127.0.0.1] (localhost [127.0.0.1]) by thinkpad.local (Postfix) with ESMTP id 32BB3AC044E; Thu, 27 Apr 2017 18:09:39 +0100 (BST) Subject: Re: svn commit: r439421 - in head/www/nginx: . files To: Ian Lepore , "Sergey A. Osokin" , Mathieu Arnold References: <201704252318.v3PNI9Io032993@repo.freebsd.org> <20170426165333.GC18747@FreeBSD.org> <1426aa14-e30f-bb33-6e76-5fdfdd222e74@FreeBSD.org> <20170427164627.GE18747@FreeBSD.org> <1493312147.66427.6.camel@freebsd.org> Cc: ports-committers@freebsd.org, svn-ports-all@freebsd.org, svn-ports-head@freebsd.org From: Vsevolod Stakhov Message-ID: <1fb15b0d-db5c-3707-6821-50cc4f57cb8e@FreeBSD.org> Date: Thu, 27 Apr 2017 18:09:39 +0100 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.8.0 MIME-Version: 1.0 In-Reply-To: <1493312147.66427.6.camel@freebsd.org> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 8bit X-BeenThere: svn-ports-head@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: SVN commit messages for the ports tree for head List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 27 Apr 2017 17:09:52 -0000 On 27/04/17 17:55, Ian Lepore wrote: > On Thu, 2017-04-27 at 16:46 +0000, Sergey A. Osokin wrote: >> On Thu, Apr 27, 2017 at 08:24:55AM +0200, Mathieu Arnold wrote: >>> >>> Le 26/04/2017 ?? 18:53, Sergey A. Osokin a ??crit : >>>> >>>> On Wed, Apr 26, 2017 at 06:18:44PM +0200, Mathieu Arnold wrote: >>>>> >>>>> Le 26/04/2017 ?? 01:18, Sergey A. Osokin a ??crit : >>>>>> >>>>>> Author: osa >>>>>> Date: Tue Apr 25 23:18:09 2017 >>>>>> New Revision: 439421 >>>>>> URL: https://svnweb.freebsd.org/changeset/ports/439421 >>>>>> >>>>>> Log: >>>>>> Upgrade from 1.10.3 to 1.12.0. >>>>>> >>>>>> ChangeLog: http://nginx.org/en/CHANGES-1.12 >>>>>> >>>>>> Remove IPV6 knob, IPv6 now compiled-in automatically if >>>>>> support is found. >>>>> This still feels like a very bad idea. >>>> Could you please explain what exactly is very bad here? >>> We had that talk like a week or two ago. It means that if the >>> package >>> builder support IPv6 it will not work on a box without it, and if >>> the >>> package builder does not support IPv6, a box with IPv6 will not be >>> able >>> to use nginx with IPv6. >> Then you can go ahead and enable IPv6 on the package builder like you >> did the >> same for the third-party moz_zip module. >> >> INET6 in FreeBSD's GENERIC kernel for years, I see no reason why >> shouldn't use it. >> > > Why do you continue to argue with multiple people who've expressed a > real-world need for this utterly trivial request to leave the IPV6 knob > in place? > > In the real world I have builder machines which DO have IPv6 enabled, > which must be able to create packages that run on machines that do NOT > have IPv6 enabled. Despite of the fact that I hate the most of IPv6 implementations, I really *hate* when somebody tries to complicate it even more. It is 2017 year so far and almost 20 years since IPv6 has been formally introduced. Why do we still support IPv6 less configurations anywhere? Perhaps, we should go further and start support IPv4 less configurations (just IPX like in 90s) or TCP less configurations? These things should just die. Come on, 20 years is enough time frame to state that IPv6 must be supported by any sane OS. I'd vote to remove this legacy knob from the ports completely (and that will simplify pkg job as well).