Return-Path: Received: from WX-Edge.insigma.com.cn (mx.insigma.com.cn [115.236.48.170]) by yocto-www.yoctoproject.org (Postfix) with ESMTP id 0D88BE006C7 for ; Fri, 30 Mar 2012 00:37:35 -0700 (PDT) Received: from WX-EXCHMB2.insigma.com.cn (10.0.1.203) by WX-Edge.insigma.com.cn (10.0.1.205) with Microsoft SMTP Server (TLS) id 8.3.83.0; Fri, 30 Mar 2012 15:36:36 +0800 Received: from [192.168.0.98] (122.224.76.38) by Wx-ExchMB2.insigma.com.cn (10.0.1.203) with Microsoft SMTP Server id 8.3.83.0; Fri, 30 Mar 2012 15:37:10 +0800 Message-ID: <1333092897.10988.5.camel@localhost.localdomain> From: Ni Qingliang To: Giannis Damigos Date: Fri, 30 Mar 2012 15:34:57 +0800 In-Reply-To: References: <1332815600.25277.22.camel@localhost.localdomain> <4F717F1A.8090108@communistcode.co.uk> <1332842944.25277.42.camel@localhost.localdomain> <1332926479.25277.66.camel@localhost.localdomain> <4F72E29E.7000803@communistcode.co.uk> <1332985943.25277.85.camel@localhost.localdomain> <1332986753.25277.87.camel@localhost.localdomain> <1333007302.25277.114.camel@localhost.localdomain> <1333012321.25277.127.camel@localhost.localdomain> <1333089019.10988.1.camel@localhost.localdomain> Organization: Insigma X-Mailer: Evolution 3.2.3 MIME-Version: 1.0 Cc: "yocto@yoctoproject.org" Subject: Re: who are using archlinux? X-BeenThere: yocto@yoctoproject.org X-Mailman-Version: 2.1.13 Precedence: list Reply-To: niqingliang@insigma.com.cn List-Id: Discussion of all things Yocto List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 30 Mar 2012 07:37:36 -0000 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 8bit On Fri, 2012-03-30 at 15:23 +0800, Giannis Damigos wrote: > On Fri, Mar 30, 2012 at 9:30 AM, Ni Qingliang > wrote: > > maybe we should report the two bugs? > > > > On Thu, 2012-03-29 at 17:12 +0800, 倪庆亮 wrote: > >> yes, you are right. > >> > >> the 32 bit lib and gconf is TWO problem, but maybe the reason is same, > >> the 'ld' search path. > >> > >> the 32 bit lib is only occured on NOT pure 64 bit arch. if you are using > >> 64 bit arch, and with this problem, maybe you have installed some 32 bit > >> libs. (my VM is a pure 64 bit arch, no this problem, and my host is > >> using multilib, with this problem) > >> > >> but the gconf problem is occured both on pure/NOT-pure 64 bit arch. > >> > >> > >> the former is focused on the arch (is 64 bit or 32 bit) of libs. and the > >> other one is focused on the dependency of libs. > >> > >> the gconf need libgtk/libgdk (found in the sysroot), and libgtk/libgdk > >> depends on libXrandr (found in host's rootfs), and libXrandr need glibc > >> 2.14 (not found in the sysroot, only glibc 2.13). > >> > >> so the search path is switch between the sysroot and host's rootfs back > >> and forth. > >> > >> On Thu, 2012-03-29 at 16:54 +0800, Giannis Damigos wrote: > >> > Installing package lib32-libxrandr, cleaning gconf and building it > >> > again does not solved the problem. > >> > Such problems on OE were solved by fixing the library path. > >> > Now gconf is looking for libXrandr in archlinux libraries and not in > >> > yocto's libraries in build directory. > >> > > >> > > >> > On Thu, Mar 29, 2012 at 10:48 AM, Ni Qingliang > >> > wrote: > >> > > give you my hands. I have waited you so long time.:) > >> > > > >> > > indeed, I have the gcc-multilib problem also. > >> > > the ati driver catalyst-total need some 32 bit libs, once you installed > >> > > that, the poky compile fail, and then we need install other 32 bit libs > >> > > (like gcc-multilib you said). (with a pure 64 bit archlinux, all are OK) > >> > > > >> > > the only reason I doubt the 'ld', only because if I remove > >> > > libXrandr.so.2 in host's rootfs, the error info changed, the libXrandr > >> > > become another lib's name (maybe libXext, I can't remember clearly). > >> > > > >> > > and after several days experiments, I can ensure the problem is focused > >> > > on the 'ld' (belongs to binutils). > >> > > > >> > > On Thu, 2012-03-29 at 15:40 +0800, Giannis Damigos wrote: > >> > >> Hi, > >> > >> > >> > >> I had similar errors building an OE image for my Micro2440 board under > >> > >> archlinux x86-64. I solved the problem by installing gcc-multlib > >> > >> (https://wiki.archlinux.org/index.php/Arch64_FAQ#Multilib_Repository_-_Multilib_Project). > >> > >> > >> > >> But, I tried to bake gconf with yocto just adding MACHINE and DISTRO > >> > >> to my local.conf: > >> > >> MACHINE ?= "qemux86-64" > >> > >> DISTRO ?= "poky-lsb" > >> > >> > >> > >> and I had the same error. > >> > >> > >> > >> On Thu, Mar 29, 2012 at 5:05 AM, Ni Qingliang > >> > >> wrote: > >> > >> > > >> > >> > Oh, I lost something, before running the testgconf.sh, you should > >> > >> > cleanall gconf, and build gconf. If not, you can't run it. > >> > >> > > >> > >> > On Thu, 2012-03-29 at 09:52 +0800, 倪庆亮 wrote: > >> > >> > > what the custom script has done is only adding the build dir (which > >> > >> > > include the 'python' symlink) into the 'PATH', only that. > >> > >> > > AND: integrate the oe-init-build-env and bitbake together. > >> > >> > > indeed, no modification. > >> > >> > > > >> > >> > > using it, I can build anything with one command (call the script), > >> > >> > > without it, I must call init env script manually. > >> > >> > > > >> > >> > > archlinux: latest > >> > >> > > poky: latest > >> > >> > > build: clean (> 4times) > >> > >> > > > >> > >> > > indeed, the possibility you mentioned has been excluded by my test. > >> > >> > > you can build it succcessfully. what arch of arch you are using? x86 or > >> > >> > > x86-64? what about the local.conf? are you using the same as mine?. > >> > >> > > > >> > >> > > both of them from my side is x86-64. > >> > >> > > > >> > >> > > I doubt 'ld', only because of my test. the attachment 'testgconf.sh' > >> > >> > > contain the root cause of the problem. and I attached the local.conf I > >> > >> > > have used, just remove the 'SOURCE_MIRROR_URL' if test it. > >> > >> > > > >> > >> > > you can run it to check the output, before that, you should change the > >> > >> > > two var in the head of it: > >> > >> > > POKY_DIR="/media/pangu/optimus" > >> > >> > > COMPILE_DIR="/media/pangu/lsbt" > >> > >> > > > >> > >> > > in it, the modification I have done (modify 'PATH') has been removed. > >> > >> > > > >> > >> > > maybe you can run it success, but you can check the 'libXrandr.so.2' > >> > >> > > line in the output, if it like: > >> > >> > > found libXrandr.so.2 at /usr/lib/../lib/libXrandr.so.2 > >> > >> > > > >> > >> > > that is the problem, even you success, it has problem. > >> > >> > > you can paste your output in the email. let's check it. > >> > >> > > the version of libxrandr on my host is 1.3.2-2. > >> > >> > > > >> > >> > > > >> > >> > > On Wed, 2012-03-28 at 18:06 +0800, Jack Mitchell wrote: > >> > >> > > > On 28/03/12 10:21, Ni Qingliang wrote: > >> > >> > > > > no echo again? > >> > >> > > > > > >> > >> > > > > what's the problem? > >> > >> > > > > > >> > >> > > > > am I wrong? where? > >> > >> > > > > > >> > >> > > > > On Tue, 2012-03-27 at 18:09 +0800, 倪庆亮 wrote: > >> > >> > > > >> thanks your reply, indeed, this is not the first email thread about > >> > >> > > > >> the > >> > >> > > > >> gconf compile problem. it in there only because of no echo. > >> > >> > > > >> > >> > >> > > > >> I have installed python2, and made a symlink in the build > >> > >> > > > >> directory, and > >> > >> > > > >> exported to the env var PATH. > >> > >> > > > >> > >> > >> > > > >> this is the script I used to build (located in the build > >> > >> > > > >> directory): > >> > >> > > > >> #!/bin/bash > >> > >> > > > >> loc_dir="`pwd`" > >> > >> > > > >> . ${loc_dir}/../optimus/oe-init-build-env ${loc_dir} > >> > >> > > > >> export PATH="${loc_dir}:$PATH" > >> > >> > > > >> bitbake $@ > >> > >> > > > >> > >> > >> > > > >> OS: x86-64 arch / desktop env: gnome 3.2 > >> > >> > > > >> local.conf ( modified section ): > >> > >> > > > >> MACHINE ??= "qemux86-64" > >> > >> > > > >> DISTRO ?= "poky-lsb" > >> > >> > > > >> SOURCE_MIRROR_URL = "http://10.10.20.149/optimussources/" > >> > >> > > > >> INHERIT += "own-mirrors" > >> > >> > > > >> BB_GENERATE_MIRROR_TARBALLS = "0" > >> > >> > > > >> INHERIT += "rm_work" > >> > >> > > > >> CCACHE = "" > >> > >> > > > >> > >> > >> > > > >> when build gconf, got error: > >> > >> > > > >> /usr/lib/../lib/libXrandr.so.2: undefined reference to > >> > >> > > > >> `memcpy@GLIBC_2.14' > >> > >> > > > >> when 'do_compile' at sanity-check step. > >> > >> > > > >> > >> > >> > > > >> the log of do_compile is attached, here is the last cmd (only > >> > >> > > > >> reserve > >> > >> > > > >> rpath): > >> > >> > > > >> ../x86_64-poky-linux-libtool ... -Wl,-rpath -Wl,/usr/lib/../lib > >> > >> > > > >> > >> > >> > > > >> the libXrandr.so.2 (depended by the libgtk/libgdk) not appear in > >> > >> > > > >> the > >> > >> > > > >> command explicitly. > >> > >> > > > >> > >> > >> > > > >> I have add --verbose in the command, and found the ld find it in > >> > >> > > > >> the > >> > >> > > > >> host's rootfs, but it should search it in the sysroot arg. > >> > >> > > > >> > >> > >> > > > >> If remove the rpath arg or with the rpath arg prefixed by sysroot > >> > >> > > > >> arg, I > >> > >> > > > >> can build success. (wich will search the libXrandr in the sysroot > >> > >> > > > >> arg > >> > >> > > > >> correctly) > >> > >> > > > >> > >> > >> > > > >> so I think the error is focused on the `ld' which should search in > >> > >> > > > >> the > >> > >> > > > >> sysroot arg at first for implicit dependended libs, but not host's > >> > >> > > > >> rootfs. > >> > >> > > > >> > >> > >> > > > >> > >> > >> > > > >> > >> > >> > > > >> On Tue, 2012-03-27 at 16:49 +0800, Jack Mitchell wrote: > >> > >> > > > >>> On 27/03/12 03:33, Ni Qingliang wrote: > >> > >> > > > >>>> who are using archlinux? I want to talk about gconf compile fail > >> > >> > > > >>>> problem. > >> > >> > > > >>>> > >> > >> > > > >>>> > >> > >> > > > >>> I use Archlinux for OE/Yocto, could you post your error so we can > >> > >> > > > >>> get an > >> > >> > > > >>> idea of the problem you are having? > >> > >> > > > >>> > >> > >> > > > >>> One note for Archlinux is that it ships Python 3 as default, to > >> > >> > > > >>> you must > >> > >> > > > >>> install Python2 from the repositories then symlink python2 to > >> > >> > > > >>> python in > >> > >> > > > >>> /usb/bin. > >> > >> > > > >>> > >> > >> > > > >>> Regards, > >> > >> > > > >>> > >> > >> > > > > >> > >> > > > I'm not 100% sure what the issue is. How come you are using a custom > >> > >> > > > build script and not the one bundled with OE? I'm not an expert by any > >> > >> > > > stretch of the imagination but by not using the given initialisation > >> > >> > > > scripts leaves me completely out of ideas as you're running a custom > >> > >> > > > setup which I couldn't possibly know the extent of. > >> > >> > > > > >> > >> > > > All I can really comment on is that Yocto builds fine and consistently > >> > >> > > > with Archlinux using the latest packages (I update nearly everyday) so > >> > >> > > > the only issue I can see is that your build setup script doesn't cut > >> > >> > > > the > >> > >> > > > mustard. Have you tried a completely clean build with only the bare > >> > >> > > > git > >> > >> > > > clone and no modifications? > >> > >> > > > > >> > >> > > > Regards, > >> > >> > > > > >> > >> > > > >> > >> > > >> > >> > -- > >> > >> > Yi Qingliang > >> > >> > niqingliang@insigma.com.cn > >> > >> > https://niqingliang2003.wordpress.com > >> > >> > > >> > >> > _______________________________________________ > >> > >> > yocto mailing list > >> > >> > yocto@yoctoproject.org > >> > >> > https://lists.yoctoproject.org/listinfo/yocto > >> > > > >> > > -- > >> > > Yi Qingliang > >> > > niqingliang@insigma.com.cn > >> > > https://niqingliang2003.wordpress.com > >> > > > >> > > > > -- > > Yi Qingliang > > niqingliang@insigma.com.cn > > https://niqingliang2003.wordpress.com > > > > First, we should test if the problem happens to an other x86-64 linux > distro or i686 Archlinux installation. > I have to setup a VM because I only use Arch. > > PS: > I read the Mailing List Guidelines from the wiki and they suggest not > to top post and to use interleaved (bottom) posting. I should have > read it before my first reply. the same problem on my side.:) -- Yi Qingliang niqingliang@insigma.com.cn https://niqingliang2003.wordpress.com