From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from lists.gentoo.org (pigeon.gentoo.org [208.92.234.80]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by finch.gentoo.org (Postfix) with ESMTPS id 32A31139337 for ; Mon, 26 Jul 2021 20:43:37 +0000 (UTC) Received: from pigeon.gentoo.org (localhost [127.0.0.1]) by pigeon.gentoo.org (Postfix) with SMTP id E6728E095F; Mon, 26 Jul 2021 20:43:32 +0000 (UTC) Received: from smtp.gentoo.org (smtp.gentoo.org [140.211.166.183]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by pigeon.gentoo.org (Postfix) with ESMTPS id 9DFFDE0940 for ; Mon, 26 Jul 2021 20:43:32 +0000 (UTC) Message-ID: <020ae083e8f14107626ad8c87d7813e383a61cac.camel@gentoo.org> Subject: Re: [gentoo-dev] Guarantees of unstable architectures From: =?UTF-8?Q?Micha=C5=82_G=C3=B3rny?= To: gentoo-dev@lists.gentoo.org Date: Mon, 26 Jul 2021 22:43:27 +0200 In-Reply-To: <57d9469d-ed7f-3641-b2c1-2c2916e53c6d@gentoo.org> References: <57d9469d-ed7f-3641-b2c1-2c2916e53c6d@gentoo.org> Organization: Gentoo Content-Type: text/plain; charset="UTF-8" User-Agent: Evolution 3.40.3 Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-Id: Gentoo Linux mail X-BeenThere: gentoo-dev@lists.gentoo.org Reply-to: gentoo-dev@lists.gentoo.org X-Auto-Response-Suppress: DR, RN, NRN, OOF, AutoReply MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Archives-Salt: 131ce92b-1cf1-451b-9a4c-7a54658c7756 X-Archives-Hash: e986381d8e634a872e475413dc7e1c53 On Mon, 2021-07-26 at 17:23 +0100, Marek Szuba wrote: > Dear everyone, > > During the open-floor part of this month's Council meeting I asked > whether there is any official policy regarding what is or is not > guaranteed for hardware architectures we do not consider stable in > Gentoo. For reference, according to the current version of > profiles/arches.desc (commit 7bdebec50c44c0222bf76334c34926b593e94dd4, > dated 2021-04-05) this means: alpha, ia64, m68k, mips, riscv, s390, > and all Prefix arches. For a start, your list includes architectures that have profiles with various levels of stability. Rules for stable/dev profiles are different than rules for exp profiles, and therefore all architectures that do not have stable/dev profiles give lower stability guarantees. As for the remaining architectures, I don't think the rules for architecture that doesn't have stable keywords should be different than rules for ~arch packages on an architecture with stable keywords. > As it turns out, we do not in fact have any such policy. On the other > hand, during my time as a Gentoo developer I have heard from other > developers a fairly wide range of opinions on the subject - from > insisting on clean QA results, passing tests etc. regardless of whether > an arch is stable or not to assuming we guarantee nothing for unstable > arches. > > Anyway, it has been decided that it makes sense to discuss this on the > mailing list before making it a Council matter. Therefore - what do you > all think here? I think the rough rule of thumb should be: 1. For stable keywords, we try really hard not to break anything. When doing somewhat risky stuff, we drop keywords to ~arch. We generally try to test stuff properly before things go stable. 2. For ~arch keywords, we don't guarantee things won't break but we also don't break them deliberately. When doing very risky stuff, we use masks or drop keywords entirely. Breakage can still sneak in. 3. For pure exp architectures, we don't guarantee any stability. We can drop keywords or break depgraph. -- Best regards, Michał Górny