**** BEGIN LOGGING AT Mon Nov 14 02:59:57 2011 Nov 14 09:01:40 quit Nov 14 10:28:30 hi all Nov 14 10:29:15 morning Nov 14 10:35:14 hey all Nov 14 10:44:22 hi bluelightning Nov 14 10:45:06 I was experimenting with my powerpc build yesterday, it seems eglibc 2.12 (angstroms preferred version) does not build for powerpc, 2.14 did work Nov 14 12:39:54 is there someone who has use spice client Nov 14 12:59:49 hi ericben Nov 14 13:42:57 guile-native fails Nov 14 13:42:58 http://pastebin.com/9VNap3kr Nov 14 13:43:00 ideas? Nov 14 13:45:19 hmmm no Nov 14 13:45:33 the recipe doesn't seem to cp Nov 14 13:45:45 yeah Nov 14 13:45:50 used to work .... Nov 14 13:46:06 but... Nov 14 13:46:13 inherit gettext seem to be lacking Nov 14 13:46:31 or is that only for target Nov 14 13:47:19 inherit gettext is in .inc file Nov 14 13:47:55 no Nov 14 13:48:00 not in native inc Nov 14 13:50:48 hmm, that solves the build issue Nov 14 13:51:04 nice Nov 14 15:05:06 hi, there is a c++ header doing that: Nov 14 15:05:07 #include Nov 14 15:05:19 which results in ./includes/gnutls/gnutlsxx.h:4:21: fatal error: exception: No such file or directory Nov 14 15:05:55 it's for guntls Nov 14 15:06:40 does anyone knows a fix? Nov 14 15:14:54 GNUtoo: not correctly build g++ cross toolchain or path to the directory where exception is in is missing Nov 14 15:22:05 ok Nov 14 15:22:25 I tought that some dep of C++ toolchain was missing Nov 14 15:22:27 I'll look Nov 14 15:22:29 thanks a lot Nov 14 15:26:22 I'll clean the toolchain and retry Nov 14 16:27:30 is anyone here working with the marvel armada 510? i am looking to see if anyone did a machine conf file for it yet. Nov 14 16:30:00 geiseri, in oe-classic or in oe-core? Nov 14 16:30:18 did anyone got that error: Nov 14 16:30:28 im in oe classic, but if its oe-core i can use that too Nov 14 16:30:31 undefined reference to `floorf@GLIBCXX_3.4.3' Nov 14 16:30:44 in oe classic look in conf/machines/ Nov 14 16:30:49 or something like that Nov 14 16:30:55 it'll tell you what is there Nov 14 16:31:28 i didnt see it in classic... or at least nothing obviously named. there was just kirkwood. ill check oe-core though. Nov 14 16:31:48 so it's not supported yet Nov 14 16:31:56 to add a machine you need: Nov 14 16:32:08 machine config + kernel+ kernel config Nov 14 16:32:19 that's the minimum I think Nov 14 16:32:20 are you guys talking about sheevaplug by any chance? Nov 14 16:32:22 okay, its in ubuntu, so i can get hte kernel, and uboot as well as the configs Nov 14 16:32:30 bluelightning: the d2plug Nov 14 16:32:36 no idea Nov 14 16:32:40 I don't think so Nov 14 16:32:41 ah ok Nov 14 16:32:41 ah right, I'm not familiar with that Nov 14 16:32:43 dream plug? Nov 14 16:33:08 http://www.plugcomputer.org/development-kits/d2plug.html <- i think it uses a different chip. Nov 14 16:33:30 cool :) Nov 14 16:33:48 * geiseri has no clue about it... i was just curious. Nov 14 16:34:20 * geiseri will pillage the ubuntu tree and see if i can make a machine config for it Nov 14 16:34:43 who should I ask for floor issue? Nov 14 16:34:47 JaMa|off, ? Nov 14 16:38:24 GNUtoo: missing -lm? Nov 14 16:39:55 hmmm I'll look Nov 14 16:40:02 thanks a lot for the help Nov 14 16:40:53 JaMa|off, lm is there Nov 14 16:41:59 basically I had an issue with c++ Nov 14 16:42:08 then I recompiled the toolchain Nov 14 16:42:15 and that issue seem gone Nov 14 16:42:21 but floor issue appeared Nov 14 16:42:35 (undefined reference to `floorf@GLIBCXX_3.4.3') Nov 14 16:42:45 not only on gtk but it seem on other packages too Nov 14 16:53:39 has anyone observed bitbake 1.13.3 (oe core) more or less randomly rebuildin dependencies? Nov 14 16:54:18 sometimes I want to rebuild one package, I know all its deps have been bult, I clean only this package and build it again, but that trigger a rebuild of couple of other packages that I did not touch Nov 14 16:54:43 I'm not 100% sure though, it's kind of strange Nov 14 16:55:01 and it seems to always want to rebuild u-boot when I build an image Nov 14 16:55:09 a rootfs image that is Nov 14 16:56:33 I've seen a situation where u-boot is rebuilt every time. if you have two recipes that both use the same git repo with different SRCREVs but whose srcrevs are based on the same branch, the two share the same auto incrementing count in the persistent data, so it flip flops between them Nov 14 16:56:42 not sure if thats your issue or not, just one i've seen before Nov 14 16:56:44 it's quite odd. Nov 14 16:57:23 I'll keep an eye on this, maybe I missed something Nov 14 17:02:46 kergoth: btw regarding uboot, is there any way around this, i.e. if I really suffer from the problem that you described, is there anything in the recipe that I can set to get around this? Nov 14 17:04:06 haven't found anything yet, but it's been a minor annoyance only, so it's not a priority for me to dig into just yet Nov 14 17:10:32 has anyone made something like an idiots guide to making your own feed repository? Nov 14 17:10:51 gchiii: for rpm or ipk? Nov 14 17:11:00 ipk Nov 14 17:11:05 don't think so, but there isn't really anything to it. take the output of deploy/ipk (or whatever), publish it, job done. Nov 14 17:11:26 there's also a "package-index" target I think for generating the index Nov 14 17:11:27 you can configure apache to look directly inside tmp/deploy if it's just for internal use, or do something a bit more robust for wider deployment. Nov 14 17:12:08 i.e. bitbake package-index Nov 14 17:13:15 yeah, my customer wants their "distro" to be angstrom-like, without being angstrom. so the feed configuration looks like angstrom, but they want the repo on their own website, but they don't want me to spend much "time" on it Nov 14 17:15:19 gchiii: I think you'd have to set ANGSTROM_FEED_CONFIGS to the name of the package that provides your feed configuration Nov 14 17:15:43 looks like another kaeilos :) Nov 14 17:17:24 ok, maybe instead of trying to figure out the way to use the "upload-packages.sh" and "sort.sh" from the contrib directory i should just make a simple feed package that expects everything to be in one place, and copy the deploy/ipk directory there. Nov 14 17:17:39 gchiii: and my feed config looks like that https://gitorious.digitalstrom.org/dss-oe/dss-oe/blobs/master/targets/dss11-devel/recipes/dss11-feeds/dss11-devel-feeds_0.1.bb Nov 14 17:17:40 hope that helps Nov 14 17:17:58 thanks Nov 14 18:49:47 kergoth: I just checked, my uboot is using a tarball, so no srcrev pingponging Nov 14 18:49:58 I wonder why it is being rebuilt each time I build an image? Nov 14 18:50:19 no clue. try bitbake -e to check its PV/PR/PE each time? Nov 14 18:50:38 or do you mean built rather than used from sstate when using an empty tmpdir? two different situations Nov 14 18:50:55 I am doing incremental builds, so not cleaning tmp Nov 14 18:51:01 * kergoth_ nods Nov 14 18:51:15 check pv/pr/pe.. only that and PN go into the stamp files, so that should be the only thing to cause it to rebuild Nov 14 18:51:15 I may change some package in the image (bump PR) and do a bitbake myimage Nov 14 18:51:23 * kergoth_ nods Nov 14 18:51:31 naturally this somepackage that is needed for the image is being rebuilt, but hten u-boot is being rebuilt too Nov 14 18:51:38 although somepackage is unrelated to uboot Nov 14 18:51:53 hmm ok let me see Nov 14 18:52:10 it's a u-boot_2009.06.bb from oe-classic Nov 14 18:52:24 maybe something goes wrong because of the wierdo-versioning Nov 14 18:52:39 it becomes u-boot-2009.06-r2 Nov 14 18:53:31 nothing wrong with that Nov 14 18:53:47 hmm indeed Nov 14 18:53:53 should I -e myimage or uboot? Nov 14 18:56:41 both and compare Nov 14 18:57:14 s/both/I would do/ Nov 14 18:57:44 checking now.. Nov 14 19:03:40 well hmm, PV="2009.06", PR="r2", can't find PE... and PN is PN="u-boot" Nov 14 19:03:51 that all look OK Nov 14 19:04:14 I am not quite sure what I should be looking for.. Nov 14 19:05:54 the image -e log contains well, the version from the image Nov 14 19:13:28 hi, did anyone had that: Nov 14 19:14:14 undefined reference to `floorf@GLIBCXX_3.4.3' Nov 14 19:14:28 how do I rebuild the toolchain for preventing that? Nov 14 19:14:42 what should I rebuild beside the follwoing Nov 14 19:15:17 GNUtoo: just guessing.. could it be some soft vs hard float configuration issue? Nov 14 19:15:41 gcc gcc-cross gcc-cross-initial gcc-cross-intermediate gcc-runtime binutils binutils-cross eglibc eglibc-initial Nov 14 19:15:55 possible let me look Nov 14 19:16:26 I saw some pastebin of that Nov 14 19:16:32 but nothing more Nov 14 19:16:38 GNUtoo: I have seen it few times: http://www.mail-archive.com/shr-user@lists.shr-project.org/msg05651.html Nov 14 19:16:38 maybe I didn't enter the right search term Nov 14 19:17:13 ok but here it's at compile time Nov 14 19:19:26 i see issues like that when using native sstate-cache from a newer distro on an older distro Nov 14 19:22:08 kergoth: now I go again.. updated my lighttpd recipe, bitbaking the image... and I get NOTE: package u-boot-2009.06-r2: task do_unpack: Succeeded and so on.. Nov 14 19:22:47 did not see anything suspicious in -e Nov 14 19:26:25 JaMa|off, I've read all the thread but since it's at compile time I cannot opkg install --force-reinstall Nov 14 19:26:36 for instance gtk+ Nov 14 19:26:45 I'll look for the hard float/soft float Nov 14 19:31:22 the target is armv7 Nov 14 19:31:24 so I have: Nov 14 19:31:26 TARGET_FPU = "vfp-neon" Nov 14 19:31:59 * Jin^eLD has never built anyting for armv7, no idea... Nov 14 19:33:38 ok thanks for the help anyway Nov 14 19:39:28 03Todd Kempel  07master * ree6171aee5 10openembedded.git/recipes/batctl/ (batctl.inc batctl_2011.3.0.bb): (log message trimmed) Nov 14 19:39:28 batctl: Add version 2011.3.0 (initial recipe) Nov 14 19:39:28 * batmand is included in the current OE repository but is not being actively maintained. Nov 14 19:39:28 The batman-adv kernel module is a replacement for batmand and requires batctl for operation. Nov 14 19:39:28 Recipe has been tested and binary found fully operational using Ȧngström on gumstix board. Nov 14 19:39:28 Signed-off-by: Todd Kempel Nov 14 19:39:29 Correct license from GPL to GPLv2+. Nov 14 19:51:17 what's the command to skip the insane checks? Nov 14 19:53:43 hi, i know this isnt really an OE question more an embedded linux question (sorry but don't know a channel for eLinux). When I boot a target from a ramdisk md autodetect isn't called. is this a known issue? Nov 14 20:04:01 http://pastebin.com/5PGX6MZr Nov 14 20:04:07 any ideas on how to work around this Nov 14 20:04:21 th emessage is confusing, the path it complains about looks ok Nov 14 20:08:53 it thinks that this path is the host include path? weird indeed Nov 14 20:09:06 cpo, there is #elinux I guess Nov 14 20:09:25 and #edev Nov 14 20:16:23 any recommendations for supporting touchscreen interface? should I set up tslib or something else? Nov 14 20:17:05 GNUtoo/florian thx Nov 14 20:17:51 gchiii, do you want Xorg? Nov 14 20:18:19 yes, i want xorg Nov 14 20:18:29 gchiii: i use evtouch with xorg. works like a charm Nov 14 20:19:07 is it difficult to setup and calibrate? Nov 14 20:19:24 gchiii: not at all Nov 14 20:20:43 ok, i see a recipe for xf86-input-evtouch. do i need anything else? Nov 14 20:23:07 a configured xserver Nov 14 20:25:04 i never know what should be in an xserver configuration, but I guess its about time I learned. Nov 14 20:27:15 gchiii, then try evdev + xinput_calibrator if needed Nov 14 20:27:36 if you have an evdev interface Nov 14 20:36:49 am I going to have to get xinit working for the 2011.03-maintenance branch? Nov 14 21:00:51 bother Nov 14 21:01:00 depmod-3.0 is brokenin classic now Nov 14 22:03:11 I just did a build with OE for the first time Nov 14 22:04:07 The kernel image might be ok. But do I also have to load a new filesystem? I had initially tested the board by putting on bootstrap, u-boot, kernel, filesystem, all of which were pre-built Nov 14 22:04:32 So now I replaced the pre-built kernel image with the one I just built. Nov 14 22:05:16 Do I need to change the fs? I get this message over the serial port: Starting udevKernel panic - not syncing: Attempted to kill init! Nov 14 22:20:52 anybody done this before? Nov 14 23:47:47 http://pastebin.com/5PGX6MZr Nov 14 23:51:22 Crofton|work: hmm, doesn't look right... **** ENDING LOGGING AT Tue Nov 15 02:59:57 2011