Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from aws-us-west-2-korg-lkml-1.web.codeaurora.org (localhost.localdomain [127.0.0.1]) by smtp.lore.kernel.org (Postfix) with ESMTP id 5B419CFA754 for ; Mon, 7 Oct 2024 21:08:59 +0000 (UTC) Received: from mail-ej1-f41.google.com (mail-ej1-f41.google.com [209.85.218.41]) by mx.groups.io with SMTP id smtpd.web11.3311.1728335333376150375 for ; Mon, 07 Oct 2024 14:08:53 -0700 Authentication-Results: mx.groups.io; dkim=pass header.i=@gmail.com header.s=20230601 header.b=E8aeFbCA; spf=pass (domain: gmail.com, ip: 209.85.218.41, mailfrom: claus.stovgaard@gmail.com) Received: by mail-ej1-f41.google.com with SMTP id a640c23a62f3a-a99415adecaso370617266b.0 for ; Mon, 07 Oct 2024 14:08:53 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20230601; t=1728335332; x=1728940132; darn=lists.openembedded.org; h=mime-version:user-agent:content-transfer-encoding:references :in-reply-to:date:to:from:subject:message-id:from:to:cc:subject:date :message-id:reply-to; bh=6bGGk06VgBk1CCjzIwtEvBL9I+DodnlPPoj/YPYYUhs=; b=E8aeFbCAz7dRGgnOQRolm2ao0tXkoeIQ3ori2PgZR5GoqyzmfhN6IHuyMHFWtln7VG 0AjFm2F1GiDHjKATVMxM8U1mE7n2pzfRPYNv6cRrlu+uXWkAvMSx7rYYAFgit2y/Cfi6 UD8eksu5Fd4i841XLjnm7ZhH/9ijGfvp+byL3CbFIx8wV+h2HDRCjZBQ8Zt5xWyunZo2 70qqT6gUhZJF4EqFZ8gSNLHS8IYkU4JWtmn9GHBEtfiuYdZDSdr6llniuJrOj+qWYO4U JFUx++5Q+4b0AilK4aJBpCcsnPCNxlgV/qdx/2XSS/IZaCc3qXbYnIWNHfTGKbOgDCuM d5Lw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1728335332; x=1728940132; h=mime-version:user-agent:content-transfer-encoding:references :in-reply-to:date:to:from:subject:message-id:x-gm-message-state:from :to:cc:subject:date:message-id:reply-to; bh=6bGGk06VgBk1CCjzIwtEvBL9I+DodnlPPoj/YPYYUhs=; b=ndiRUQ1MaapJcnLd6g26bsRMixhuxl3lWFUkEmp0CJMJEcCrD5P8yGDjr8OBG4cIVu PC6ZHCQBcuYvsEss6Dy3JwwQS7/sO5Dano0EMGizeMhQGdPquFIDkjnMYJAfkITffjzV VcJDKoZN531PGXbqNOfQBYjEsuetP3V4E5yGwBBCc2NJ5hP9qSBBRW/pWLk74lqiJWEK mX3H20iJCEn5FEN0DWYQWvtNL7ZKMQTo/HYdG4MW3Ruiqi9MyQoXO3URfhIYODq7ENPc Y6ePk9DwCe8cBpo1F1FqH2O15iaU3+eqiEDn1aCtQ4Ao9gDK+ndxGCe5vTTrBRMFCPzB cbQg== X-Forwarded-Encrypted: i=1; AJvYcCWwy0Pn1OD5SRfqDNrXOeBSPHtZaYcQ/mxXldBc4S4MSEKy7fOMaAoq639KA2I3dwu7kcHQc2WTa/xpA8wT7OO9AQ==@lists.openembedded.org X-Gm-Message-State: AOJu0YzkFl5XGcO2jera28syhcFISHcT1S5EqLqU+4XglHMbYzCWvL7W Ov/e87fUal4yBrh7FzI0izScpbm+jMeb8UPD9cwomO1nhHoLoT1/ X-Google-Smtp-Source: AGHT+IFUY/O3pL+zX9Gz6ZXaNlonElogp75Qg3myyfbkUGrRzi7f5YiaoRI1iQF2ax6dLcsg54OZMw== X-Received: by 2002:a17:907:2ce3:b0:a86:9fac:6939 with SMTP id a640c23a62f3a-a99679501c6mr86852966b.30.1728335331305; Mon, 07 Oct 2024 14:08:51 -0700 (PDT) Received: from [192.168.1.70] (0x57346aa5.static.cust.fastspeed.dk. [87.52.106.165]) by smtp.gmail.com with ESMTPSA id a640c23a62f3a-a993d92ed5dsm362307866b.63.2024.10.07.14.08.50 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 07 Oct 2024 14:08:50 -0700 (PDT) Message-ID: <2456863182484d2b15f75fe0f67a3473240a3fe6.camel@gmail.com> Subject: Re: [OE-core] [RFC][master][PATCH] populate_sdk_base: Add SDKIMAGE_INSTALL_COMPLEMENTARY_HARD_DEPENDS_ONLY From: Claus Stovgaard To: Richard Purdie , openembedded-core@lists.openembedded.org, Ross.Burton@arm.com Date: Mon, 07 Oct 2024 23:08:49 +0200 In-Reply-To: References: <20240927132040.200122-1-claus.stovgaard@gmail.com> Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable User-Agent: Evolution 3.50.2 MIME-Version: 1.0 List-Id: X-Webhook-Received: from li982-79.members.linode.com [45.33.32.79] by aws-us-west-2-korg-lkml-1.web.codeaurora.org with HTTPS for ; Mon, 07 Oct 2024 21:08:59 -0000 X-Groupsio-URL: https://lists.openembedded.org/g/openembedded-core/message/205280 On Tue, 2024-10-01 at 10:53 +0100, Richard Purdie wrote: > On Fri, 2024-09-27 at 15:20 +0200, Claus Stovgaard via > lists.openembedded.org wrote: > > SDKIMAGE_INSTALL_COMPLEMENTARY_HARD_DEPENDS_ONLY is a new variable > > to > > enable installation of recommended complementary packages in the > > classic > > sdk. Use it to get populate_sdk behave like Kirkstone and earlier. > >=20 > > As installing complementary recommended packages is known to cause > > bugs > > it was disabled. Though it can be usefull when having dummy > > packages > > with DEPENDS to create SDK's. Use this variable with caution. > >=20 > > Signed-off-by: Claus Stovgaard >=20 > The commit message needs improvement since I can't tell what this > means/does. >=20 > What was the behaviour in kirkstone and how did it change? Why/when > would I need it? >=20 > What is a "hard depends"? >=20 > I don't have an opinion on whether should do something like this or > not > as yet, mainly as I can't tell what it is doing. Can you explain it > without reference to "behaviour in kirkstone"? >=20 > I also have an aversion to "1" as a magic value. Have a look at > bb.utils.to_boolean() for that. >=20 Thanks for your comments. I will try to improve the commit message, and split the documentation part out from the rest. The issue it tries to solve is the one described here. https://lists.openembedded.org/g/openembedded-core/topic/108100662#msg20374= 5 In short - with this commit https://git.yoctoproject.org/poky/patch/?id=3D4705dd264681d908f144dd4d9bf1f= 6175f68d8b9 the behaviour of complementary package handling changed ignore RRECOMMENDS. This was changed to fix some long standing bugs for RDEPENDS from -dev packages. Though as the complementary handling of including packages from RRECOMMENDS is needed in some situations the purpose of this patch is to change from hardcoded disabled to be a configurable option. A prime example for where we need the installation of packages in RRECOMMENDS, and not ignore it is for generating a SDK with all dependencies for an application. E.g. if we have a dummy recipe with packages in DEPENDS, the packages listed in DEPENDS will end in RRECOMMENDS. See process_depchains in package.py for the pkg_adddeprrecs method, and not be installed in the SDK. Using the variable introduced with this patch, will install packages from RRECOMMENDS in the SDK, and we have a classic SDK with all the packages from DEPENDS and ready to be used for the application. Hope it make better sense for you now, and I will try to describe it better in the commit message. Regards Claus > Cheers, >=20 > Richard