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 1RGoah-00061Q-Vx for openembedded-core@lists.openembedded.org; Thu, 20 Oct 2011 11:07:08 +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 1RGoV4-00056E-B1 for openembedded-core@lists.openembedded.org; Thu, 20 Oct 2011 11:01:18 +0200 From: Phil Blundell To: Patches and discussions about the oe-core layer Date: Thu, 20 Oct 2011 10:01:17 +0100 In-Reply-To: <4E9FD2B2.5000508@intel.com> References: <66cdd5054e220fcce980222ed68a2afc37f614af.1318984111.git.nitin.a.kamble@intel.com> <35417d38f61615445ed0c4d875e0bd0aa5649af4.1318984111.git.nitin.a.kamble@intel.com> <4E9FD2B2.5000508@intel.com> X-Mailer: Evolution 3.0.2- Message-ID: <1319101278.22985.258.camel@phil-desktop> Mime-Version: 1.0 Subject: Re: [PATCH 2/6] gmp: also generate the libgmpcxx library 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, 20 Oct 2011 09:07:08 -0000 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit On Thu, 2011-10-20 at 00:50 -0700, Saul Wold wrote: > The example case I found was a coreutils package that was built prior to > this change failed to fulfill it's dependencies after this change when > creating an image. If I rebuild coreutils with the newer gmp build, > then all is well, I think we need to understand this better before > making this change. > > Did something else change in the packaging that would cause use to look > the mapping above? If you've now got two DSOs in one output package (which it sounds like you might have, since libgmpcxx was apparently added) then the autonamer will be defeated. They need to go in separate PACKAGES for that to work out right. p.