From owner-svn-src-all@FreeBSD.ORG Fri Apr 10 10:03:16 2015 Return-Path: Delivered-To: svn-src-all@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id A23A9478; Fri, 10 Apr 2015 10:03:16 +0000 (UTC) Received: from mx0.deglitch.com (unknown [IPv6:2001:16d8:ff00:19d::2]) by mx1.freebsd.org (Postfix) with ESMTP id 1A2BABD9; Fri, 10 Apr 2015 10:03:16 +0000 (UTC) Received: from [192.168.11.8] (unknown [98.248.95.7]) by mx0.deglitch.com (Postfix) with ESMTPSA id 35D428FC27; Fri, 10 Apr 2015 14:03:02 +0400 (MSK) Content-Type: text/plain; charset=utf-8 Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2098\)) Subject: Re: svn commit: r281129 - head/etc From: Stanislav Sedov In-Reply-To: <20150410092942.GF99375@FreeBSD.org> Date: Fri, 10 Apr 2015 03:02:54 -0700 Content-Transfer-Encoding: quoted-printable Message-Id: <7D18BE54-E0D3-42D6-A00F-1CF7DE435DD8@freebsd.org> References: <20150406085323.GL23643@zxy.spb.ru> <20150409215802.GY21982@hades.panopticon> <20150409220442.GQ8220@zxy.spb.ru> <20150409232711.GZ21982@hades.panopticon> <20150409234026.GH1394@zxy.spb.ru> <20150410000259.GA21982@hades.panopticon> <20150410070845.GU8220@zxy.spb.ru> <20150410073230.GE99375@FreeBSD.org> <20150410092942.GF99375@FreeBSD.org> To: Gleb Smirnoff X-Mailer: Apple Mail (2.2098) Cc: "src-committers@freebsd.org" , "svn-src-all@freebsd.org" , Dmitry Marakasov , Slawa Olhovchenkov , "svn-src-head@freebsd.org" , Garrett Cooper X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.18-1 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 10 Apr 2015 10:03:16 -0000 > On Apr 10, 2015, at 2:29 AM, Gleb Smirnoff = wrote: >=20 > Yes, we do. And it is long overdue. >=20 > Today vast majority of Russian users do this change manually. > Minority stays with KOI8-R. Even Slawa mentions that he is > migrating machines. >=20 > Yes, this has impact on people doing upgrade. But before we > have impact on people doing fresh install. Since today most > Russian users already did this change manually, we expect impact > to be small. >=20 > Let me enumerate problems for those who stay with KOI8-R: >=20 > If your terminal is in KOI8: > - Emails with umlaut in From: are mangled. You compose reply, > you put "?" char into person's name. > - Emails from Japanese and Chinese people are even worse. The > entire name is ??????. >=20 > If your filesystem is in KOI8: > - Email attachments you save have names mangled. > - Many Gnome based applications crash. > - You can't read file names from downloaded torrents with Russian > file names inside, neither you can create your own torrent shares > that would be compatible with others. >=20 > Those who use FreeBSD as desktop, except very small minority, > switched to UTF-8. I=E2=80=99m not arguing the benefits of UTF8/KOI8 and userbase sizes of = both, although I don=E2=80=99t necessarily agree with all the statements = above. The question here is how to introduce the change in the least disruptive way. And I fail to see how adding a new, russian_utf login class is inferior to modifying the existing one, which is for better or for worse was using koi8-r for decades. The way I see it, if we modify the existing class: * New and existing users created without login class are unaffected; * New users created with the =E2=80=98russian=E2=80=99 class explicitly set will have UTF-8 encoding enabled; * Existing users with the =E2=80=98russian=E2=80=99 login class switch = from koi8-r to utf8. This is a no-op for the UTF8 russian userbase, and a forced change for KOI8 users as they will have to either modify the login class, or change the login class of existing users * There=E2=80=99s no KOI8 login class anymore, all KOI8 users will have to add a =E2=80=98custom=E2=80=99 KOI8 login class. If we add a new, =E2=80=98russian_utf8' login class: * New and existing users created without login class are unaffected; * New users created with the =E2=80=98russian=E2=80=99 class explicitly set will have KOI-8 encoding enabled; * New users created with the =E2=80=98russian_utf8=E2=80=99 class = explicitly set will have UTF-8 encoding enabled; * Existing russian UTF8 users are unaffected * Existing KOI8-R users are unaffected * Login classes for both UTF8 and KOI8-R are available Am I missing something?=20 -- ST4096-RIPE