From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.chez-thomas.org (hermes.mlbassoc.com [64.234.241.98]) by mx1.pokylinux.org (Postfix) with ESMTP id B2F324C80546 for ; Mon, 14 Feb 2011 17:43:47 -0600 (CST) Received: by mail.chez-thomas.org (Postfix, from userid 999) id 5BD61166027C; Mon, 14 Feb 2011 16:43:46 -0700 (MST) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on hermes.chez-thomas.org X-Spam-Level: X-Spam-Status: No, score=-2.9 required=4.0 tests=ALL_TRUSTED,BAYES_00 autolearn=ham version=3.3.1 Received: from hermes.chez-thomas.org (hermes_local [192.168.1.101]) by mail.chez-thomas.org (Postfix) with ESMTP id 7C6FA16601E2; Mon, 14 Feb 2011 16:43:44 -0700 (MST) Message-ID: <4D59BE30.9050802@mlbassoc.com> Date: Mon, 14 Feb 2011 16:43:44 -0700 From: Gary Thomas User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.13) Gecko/20101209 Fedora/3.1.7-0.35.b3pre.fc13 Thunderbird/3.1.7 MIME-Version: 1.0 To: Poky Subject: Controlling package build ordering X-BeenThere: poky@yoctoproject.org X-Mailman-Version: 2.1.13 Precedence: list List-Id: Poky build system developer discussion List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 14 Feb 2011 23:43:48 -0000 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit I have a modified version of poky-feed-config-opkg which actually builds an opkg database (as opposed to the default which does not). I've set it so that my base-feeds.conf only contains entries for valid package sets like this: for arch in $ipkgarchs; do # Only export real directories if [ -d "${DEPLOY_DIR}/ipk/${arch}" ]; then echo "src/gz ${FEEDNAMEPREFIX}-$arch ${FEEDURIPREFIX}/$arch" >> $basefeedconf fi done As you can see, only those entries which have some packages will be included. The problem I have is _when_ my feed-config-opkg.bb is actually built. If I start from scratch, it gets built pretty early and the ipk tree is often quite empty, perhaps only containing a single collection. If I rerun this at the end, say after building a complete image, then I get a complete set of entries. What I'd like to do is find a way to postpone the building of this recipe until very late, hopefully just before the final image is built. Is there any way to fine tune the build and accomplish this? Note: if I leave out the test for valid subdirectories it all works but then I get errors at runtime complaining about the empty collections. This is what I'd like to avoid. Thanks for any ideas -- ------------------------------------------------------------ Gary Thomas | Consulting for the MLB Associates | Embedded world ------------------------------------------------------------