From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [82.195.155.133] (helo=mail.techworks.ie) by linuxtogo.org with esmtp (Exim 4.69) (envelope-from ) id 1MsO08-0001S9-FL for openembedded-devel@lists.openembedded.org; Mon, 28 Sep 2009 23:43:23 +0200 Received: from archeopterix.techworks.local (unknown [86.43.120.173]) by mail.techworks.ie (Postfix) with ESMTP id E3414BC0F for ; Mon, 28 Sep 2009 22:42:39 +0100 (IST) Message-ID: <4AC12E48.3010801@techworks.ie> Date: Mon, 28 Sep 2009 22:44:40 +0100 From: Christian Gagneraud User-Agent: Thunderbird 2.0.0.21 (X11/20090320) MIME-Version: 1.0 To: openembedded-devel@lists.openembedded.org X-SA-Exim-Connect-IP: 82.195.155.133 X-SA-Exim-Mail-From: cgagneraud@techworks.ie X-SA-Exim-Version: 4.2.1 (built Wed, 25 Jun 2008 17:20:07 +0000) X-SA-Exim-Scanned: No (on linuxtogo.org); Unknown failure Subject: meta-native? X-BeenThere: openembedded-devel@lists.openembedded.org X-Mailman-Version: 2.1.11 Precedence: list Reply-To: openembedded-devel@lists.openembedded.org List-Id: Using the OpenEmbedded metadata to build Distributions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 28 Sep 2009 21:43:23 -0000 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Hi all, Is there a way to build all and only the native recipes in a single shot? Short answer I suppose: Write a meta-native recipe that depends on all the native recipes and then do a "MACHINE=native bitbake your-meta-native" But If anyone fancy explain me why my first attemp didn't give me the expected results: I've tried naively to write a meta-native recipe by putting in DEPENDS a list of native recipes (from stamps/x86_64-linux/*-native-*) that have been built after a "bitbake console-image" but still using my usual MACHINE (ts72xx). Here is my meta-native: ---- DESCRIPTION="Meta package to build all native packages in a raw" DEPENDS = "\ autoconf-native \ automake-native \ bison-native \ bzip2-full-native \ curl-native \ db-native \ dbus-glib-native \ dbus-native \ desktop-file-utils-native \ expat-native \ fakeroot-native \ file-native \ flex-native \ gdbm-native \ gettext-native \ glib-2.0-native \ gmp-native \ gnu-config-native \ gtk-doc-native \ intltool-native \ ipkg-utils-native \ libtool-native \ libxml-parser-perl-native \ libxml2-native \ m4-native \ mpfr-native \ ncurses-native \ openssl-native \ opkg-native \ pax-utils-native \ perl-native \ pkgconfig-native \ python-native \ quilt-native \ sed-native \ shared-mime-info-native \ shasum-native \ stagemanager-native \ unifdef-native \ xorg-headers-native \ xproto-native \ zlib-native \ " inherit meta ---- I expected then that a "rm -Rf $TMPDIR; MACHINE=whatever bitbake meta-native" would have build only the native recipes but it looks like some cross/target recipes are being build too, for now (the build is still ongoing) it has pulled the following "non-native" dependencies (found in stamps/armv4t-angstrom-linux-gnueabi): binutils-cross expat gcc-cross gcc-cross-initial gcc-cross-intermediate glibc glibc-initial libtool-cross linux-libc-headers Is it really a native package that pulls these dependencies, or is it because the cross-toolchain is a kind of default dependency? I'm using angstrom 2008.1-stable, my goal is to build all the native recipes and then make a big tarball of my TMPDIR to save time for the future. Ideally i would like then to build the toolchains I need, install them and use them with openembedded. If that doesn't work I will go for the big tarball as well. I'm only interesting in repeatable images creation, not repeatable toolchain and native stuff and I would like to use this big tarball as a starting point to speed-up my "build images from scratch" process. As I'm not modifying any native/cross/sdk recipes, but only recipes for the target, my way of doing should be safe to guarantee the repeatability of image building. Any thought? Regards, Chris