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 1RAk0Q-0002dL-3Z for openembedded-core@lists.openembedded.org; Mon, 03 Oct 2011 17:00:34 +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 1RAjv4-0008C9-IG for openembedded-core@lists.openembedded.org; Mon, 03 Oct 2011 16:55:02 +0200 From: Phil Blundell To: Patches and discussions about the oe-core layer Date: Mon, 03 Oct 2011 15:55:01 +0100 In-Reply-To: References: X-Mailer: Evolution 3.0.2- Message-ID: <1317653702.13337.4.camel@phil-desktop> Mime-Version: 1.0 Subject: Re: [RFC] .la files cleanup 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: Mon, 03 Oct 2011 15:00:34 -0000 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit On Mon, 2011-10-03 at 14:58 +0400, Dmitry Eremin-Solenikov wrote: > My proposal would be to include a special task or a > do_install_append() part that > will remove specified .la files. (like ${libdir}/xorg/modules/*/.la, > ${libdir}/matchbox-panel/*.la, etc.). I see this usually done in > do_install or in special do_install_append > parts, but I think it would be better to handle this globally. I guess the ideal solution would be to find a way to stop libtool from installing these things in the first place. It isn't totally obvious to me that the .la files are a net benefit even for real shared libraries, and for plugins I think they are probably a net loss in most/all cases. When this issue last came up it seemed that Richard's preferred fix was to just ship the .la files in FILES_${PN}-dev, and I guess that approach does at least have the merit of simplicity. See: http://lists.linuxtogo.org/pipermail/openembedded-core/2011-July/005533.html p.