From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail.pbcl.net ([88.198.119.4] helo=hetzner.pbcl.net) by linuxtogo.org with esmtp (Exim 4.72) (envelope-from ) id 1QS5zM-0001Xf-NO for openembedded-core@lists.openembedded.org; Thu, 02 Jun 2011 13:22:56 +0200 Received: from cambridge.roku.com ([81.142.160.137] helo=[172.30.1.145]) by hetzner.pbcl.net with esmtpsa (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.69) (envelope-from ) id 1QS5wE-00033s-AS for openembedded-core@lists.openembedded.org; Thu, 02 Jun 2011 13:19:42 +0200 From: Phil Blundell To: Patches and discussions about the oe-core layer In-Reply-To: <4DE6A9A6.4020107@windriver.com> References: <7826575ce92090c4460c7d016e0b06441f84cff7.1306865217.git.scott.a.garman@intel.com> <1306921632.2529.18.camel@phil-desktop> <4DE67A4F.5010703@intel.com> <4DE6A9A6.4020107@windriver.com> Organization: Phil Blundell Consulting Ltd Date: Thu, 02 Jun 2011 12:19:41 +0100 Message-ID: <1307013581.2529.181.camel@phil-desktop> Mime-Version: 1.0 X-Mailer: Evolution 2.30.3 Subject: Re: [PATCH 2/7] shadow: add a -native recipe with customized utilities 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: Thu, 02 Jun 2011 11:22:56 -0000 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit On Wed, 2011-06-01 at 16:05 -0500, Mark Hatle wrote: > The reason I suggested the --root option was primarily for the ease of people > who are NOT using the automated scripting, i.e. someone manually adding a > preinst (or similar) to their recipes. I was thinking about this stuff the other day for update-rc.d. Really, it sucks that all the manually-written postinsts have to include all of this garbage about checking for $D. It also sucks a bit that this same code (for both manually-written and auto-generated postinsts) ends up getting deployed into the packages and onto the target system, where it is at best a waste of space. I think maybe a better way of dealing with that whole thing would be for update-rc.d-native and similar tools to provide a set of wrapper scripts which would take care of all the machinations required to support offline root mode, be that adding "--root $D" or doing stuff with chroot or whatever. If those were installed into some appropriate place in the sysroot then we could arrange for the right directory to be prepended to $PATH during rootfs construction and then everything ought to "just work" without the postinst scripts needing to take any special measures. p.