From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mga03.intel.com ([143.182.124.21]) by linuxtogo.org with esmtp (Exim 4.72) (envelope-from ) id 1Si1JY-0006Cg-UD for openembedded-core@lists.openembedded.org; Fri, 22 Jun 2012 12:42:09 +0200 Received: from azsmga002.ch.intel.com ([10.2.17.35]) by azsmga101.ch.intel.com with ESMTP; 22 Jun 2012 03:31:20 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.71,315,1320652800"; d="scan'208";a="114990944" Received: from unknown (HELO helios.localnet) ([10.252.120.123]) by AZSMGA002.ch.intel.com with ESMTP; 22 Jun 2012 03:31:19 -0700 From: Paul Eggleton To: openembedded-core@lists.openembedded.org Date: Fri, 22 Jun 2012 11:31:18 +0100 Message-ID: <2779429.QaclkbfKlo@helios> Organization: Intel Corporation User-Agent: KMail/4.8.3 (Linux/3.2.0-25-generic-pae; KDE/4.8.3; i686; ; ) In-Reply-To: References: MIME-Version: 1.0 Subject: Re: CORE_IMAGE_EXTRA_INSTALL has potential to be misleading? X-BeenThere: openembedded-core@lists.openembedded.org X-Mailman-Version: 2.1.11 Precedence: list Reply-To: Patches and discussions about the oe-core layer List-Id: Patches and discussions about the oe-core layer List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 22 Jun 2012 10:42:09 -0000 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="us-ascii" On Wednesday 20 June 2012 14:28:40 Robert P. J. Day wrote: > was reminded of this as i was perusing some old OE core notes. > remember this from core-image.bbclass? > > CORE_IMAGE_BASE_INSTALL = '\ > task-core-boot \ > task-base-extended \ > \ > ${CORE_IMAGE_EXTRA_INSTALL} \ > ' > > CORE_IMAGE_EXTRA_INSTALL ?= "" > > IMAGE_INSTALL ?= "${CORE_IMAGE_BASE_INSTALL}" > > except there are some core image variation recipes > (core-image-minimal-dev.bb, among others) that do things like this: > > IMAGE_INSTALL = "task-core-boot ${ROOTFS_PKGMANAGE}" > > IMAGE_FEATURES += "dev-pkgs" > > IMAGE_LINGUAS = " " > > LICENSE = "MIT" > > inherit core-image > > note how the explicit assignment to IMAGE_INSTALL in that second > example will copletely override the "IMAGE_INSTALL ?=" in the bbclass > file, at which point ${CORE_IMAGE_EXTRA_INSTALL} will obviously never > be evaluated and included. Right, this is not really ideal. In the case of core-image-minimal(-dev) we're trying to be as minimal as reasonably possible, and that includes eliminating a bunch of the stuff that core-image.bbclass adds. The easiest way to do this is by simply setting IMAGE_INSTALL explicitly. It seems like core-image-minimal-dev has been a little neglected - at least two changes that went into core-image-minimal were not also applied to it, including the addition of CORE_IMAGE_EXTRA_INSTALL (or POKY_EXTRA_INSTALL as it was known at the time). I'm wondering if a better way for this specific recipe to work would just to be to "require core-image-minimal.bb" and then set IMAGE_FEATURES += "dev-pkgs" as appropriate. core-image-minimal* are somewhat special cases, however I do think we do need to improve the consistency of all of our image recipes. Cheers, Paul -- Paul Eggleton Intel Open Source Technology Centre