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 1RJork-0007Jv-SQ for openembedded-core@lists.openembedded.org; Fri, 28 Oct 2011 18:01:08 +0200 Received: from elite.brightsigndigital.co.uk ([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 4.72) (envelope-from ) id 1RJolw-0002eS-Lj; Fri, 28 Oct 2011 17:55:08 +0200 From: Phil Blundell To: McClintock Matthew-B29882 , Patches and discussions about the oe-core layer Date: Fri, 28 Oct 2011 16:54:47 +0100 In-Reply-To: References: <1319700362.15601.2.camel@ted> <1319743874-9891-1-git-send-email-msm@freescale.com> <1319792127.22423.1.camel@ted> X-Mailer: Evolution 3.0.2- Message-ID: <1319817308.22985.448.camel@phil-desktop> Mime-Version: 1.0 Subject: Re: [PATCH v2 1/2] Add new strace-graph package to avoid making perl a dependecy for all of strace 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, 28 Oct 2011 16:01:09 -0000 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit On Fri, 2011-10-28 at 15:34 +0000, McClintock Matthew-B29882 wrote: > On Fri, Oct 28, 2011 at 3:55 AM, Richard Purdie > wrote: > >> -RDEPENDS = "perl" > >> +PACKAGES_prepend = "${PN}-graph " > >> +FILES_${PN}-graph = "${bindir}/strace-graph" > >> +RDEPENDS_${PN}-graph = "perl" > > > > I should have been clearer in my original email, pretty much every other > > recipe uses += or =+ in this context. Whilst the _prepend override is > > probably ok, its probably a bit heavy in this case and the standard > > operator should work just fine. > > usr/bin/strace-graph will be picked up by the default rules for ${PN} > - by prepending I thought I was letting strace-graph get picked out > first in ${PN}-graph then everything else could go in ${PN} without > having the explicitly list a new rule for ${PN}. That logic sounds reasonable, but I think Richard's point was that it's more conventional to use "=+" rather than an override for prepending. p.