**** BEGIN LOGGING AT Fri Oct 09 03:00:00 2009 Oct 09 03:00:04 its size should not matter Oct 09 03:00:20 and by /lib i mean /usr/lib Oct 09 03:01:08 is this the first boot? maybe configuring installed packages take a lot of time Oct 09 03:01:28 two hints: a) use tcpdump/wireshark to see if NFS action take place b) boot the kernel with init=/bin/sh Oct 09 03:01:30 i don't follow Oct 09 03:01:36 like Oct 09 03:01:51 oe constructs the rootfs .tar.gz Oct 09 03:01:57 that archive never containes the file Oct 09 03:02:11 but it is in the staging tree Oct 09 03:03:49 martinmeba: well, then debug this first. If "grep" requires libintl.so then use avr32-*-objdump -x grep | less and see if it has a SO_NEEDED libintl.so. If that is the case rebuild bitbake -crebuild busbox and see if it complains that it doesn't find a provider for "libintl.so" Oct 09 03:04:10 ok Oct 09 03:04:13 i will give that a shot Oct 09 03:04:18 thanks Oct 09 06:56:09 hi all, i'm new to OE, and i'm trying to build my image for beagleboard. i checkout the 2009 stable branch and build succeed, but now i want to use the latest sgx driver which is not included in this branch ... what should i do? Oct 09 06:56:45 i also try to use the latest OE instead of 2009 stable...but i cannot build it successfully... Oct 09 06:58:43 see tinderbox results and pick a revision which has been successfully built for your board and distro? Oct 09 07:02:12 koobe: thanks...that can be great.. can you give give me more deteail? what is tinderbox anyway.. Oct 09 07:04:51 leo: I assume your talking about the closed source TI drivers, do you have the binaries from TI for a start? (note: #beagle may be a much better bet for sorting this). Oct 09 07:08:20 DJWillis:yes, i already get that bin file.. and i try to build with the latest kernel but failed, so i switch to OE which already contains a lot of patches from others...and i've try to learn OE by myself but it's not that easy...so i'm here to ask my questions :) Oct 09 07:15:42 leo: one (*cough* hacky *cough*) way to add all the .9 SGX drivers to stable/2009 would be to copy back the powervr-drivers folder from oe.dev but I can't vouch for the kernel patches that may also go with it as I tend to use said SGX drivers on the Pandora and OMAP3EVM not the Beagle. Oct 09 07:30:26 Can anyone tell me what the difference is between theese: all-angstrom-linux, i686-angstrom-linux, i686-generic-angstrom-linux, i686-linux ?? Oct 09 07:30:39 DJWillis: see...i'll try it, thanks:-) also i got some basic questions about OE, can you help with it? Oct 09 07:31:17 leo: ask away, I can't claim to no that much but there are loads of good people in the chan who can answer most things. Oct 09 07:35:56 my 1st question: i set the DISTRO = "angstrom-2008.1" and MACHINE = "beagleboard" in local.conf and if i want to add or remove the default packages, how can i do that... i checked console-image.bb and image.class but get no idea.. Oct 09 07:39:10 leo: adding is simple, add packages to the IMAGE_INSTALL list. perhaps make your own copy of console-image.bb Oct 09 07:41:18 Can you tell me where to find info regarding: all-angstrom-linux, i686-angstrom-linux, i686-generic-angstrom-linux, i686-linux ?? I'm not really interested in a distro specific explanation - just some info on what's stored where, and why... Oct 09 07:41:22 koobe: thanks. how about remove some package from the default package list of this distribution or change the package to another version. Oct 09 07:44:10 koobe:i found the tinderbox you mentioned, i can save a lot of time if i knew it earlier.. Oct 09 07:45:38 leo: what I did was first create my own image recipe which only has task-boot, then add packages from there Oct 09 07:46:50 leo: and for selecting versions of packages I've used PREFERRED_VERSION_package_name in my machine configuration, I'm not sure if that's pretty or not. Oct 09 07:47:44 Hasse__: they are the arch folders for the packages (in simple terms) Oct 09 07:50:58 DJWillis: I see the folder names and their content, but I don't see the logic in where to look for what. Oct 09 07:53:31 Hasse__: that would depend on your platform, all = all for that distro, i.e. generic stuff that is cross platform. i686-angstrom = Angstrom specific stuff for the i686 arch, etc. etc. (i686-linux only seems to get a few packages and it seems some sort of legacy hangover, never quite sure). Anyway, that should give your the idea. When your machine is something more interesting than i686 it makes a lot more sense. Oct 09 07:56:01 koobe: em...this way should work..when adding package, OE will take care with the dependency, right? Oct 09 07:56:58 koobe: and about kernel, i found PREFERRED_PROVIDER_virtual/kernel = "linux-omap3" in beagleboard.conf but there's no linux-omap3.bb under recipes.. Oct 09 07:57:09 DJWillis: Oki, thx! :) Do you know where to find even more info? Oct 09 07:57:35 Hasse__: to give you another example, armv7a-angstrom-linux-gnueabi i686-linux omap3-pandora-angstrom-linux-gnueabi Oct 09 07:57:35 is my tmp/staging layout, I have armv7a angstrom bits (for all armv7a archs), omap3-pandora bits for just that machine and the i686 (i.e. native in my case) bits that run on my build host to put it all together. Oct 09 07:58:01 Hasse__: in your case your host and target are the same so it makes it hard to see clearly Oct 09 07:58:08 koobe: for example if i want to use the kernel with PM support, how can i do that? Oct 09 07:58:10 good morning Oct 09 07:59:16 Morning mckoan Oct 09 07:59:28 leo: yes, package dependencies work nicely. Oct 09 07:59:35 DJWillis: What do you mean by same host and target - that I build on the same arch as my target? Oct 09 07:59:54 DJWillis: ... or same type of arch - i686? Oct 09 08:00:43 Hasse__: yep, your building ON an i686 box for an i686 target, in my case I build ON i686 for a OMAP3-PANDORA target (and that target has an ARMv7a CPU core). Oct 09 08:01:02 Hasse__: what do your 'think' your setup for? Oct 09 08:01:21 leo: I'd suggest you run bitbake with -D or -e options to see what's happening. Oct 09 08:01:57 morning from .de land :-D Oct 09 08:02:31 XorA|gone: .de? Not home then ;-) Oct 09 08:03:11 DJWillis: in Berlin for OE eV meeting on sunday Oct 09 08:03:44 leo: I'm not expert on how oe/bb picks the package versions, but tweaking PREFERRED_PROVIDER in machine conf and DEFAULT_PREFERENCE in recipes has worked well enough for me so far :) Oct 09 08:03:46 XorA|gone: Berlin, so Cambridge was not the location in the end? Oct 09 08:05:30 DJWillis: this is the eV seperate from OEDEM Oct 09 08:07:09 XorA|gone: for some reason I had it in my head that the special eV meeting was to be at OEDEM, I guess it makes more sense to have it in .de ;-). Oct 09 08:07:15 koobe: Got it.i think i should have a try and check by myself first, thanks for being so patient to me:) Oct 09 08:07:47 nice day guys:) Oct 09 08:11:13 morning Oct 09 08:13:31 anyway Im off to find way into town Oct 09 08:21:20 I'm using a script for setting up oe for compiling angstrom. I source this script with "source setup-oe.sh". At the end it does "cd build". What is the effect of this? The shell I did the sourcing doesn't change directory. Oct 09 08:21:56 tasslehoff: my script? Oct 09 08:23:45 good morning Oct 09 08:24:17 hi florian Oct 09 08:29:07 hrw, actually I'm merging yours and the one from http://www.angstrom-distribution.org/files/source-me.txt. It's the latter that does "cd $PKGDIR" near the end. Oct 09 08:29:39 tasslehoff: what is your shell? Oct 09 08:30:03 I'm using bash and syntax '. setup-env.sh' Oct 09 08:30:15 Jay7, bash Oct 09 08:30:23 directory changing ok Oct 09 08:31:53 Jay7, ah. that worked here as well :) Oct 09 08:32:17 :) Oct 09 08:32:36 but now, so did sourcing it. most likely pebkac :) Oct 09 08:34:01 good morning florian, all Oct 09 08:38:06 Jay7, is there a reason for doing that cd at all? If BBDIR and TMPFILES are set, it doesn't matter if you stand in the build directory, right? Oct 09 08:46:24 tasslehoff: I don't know :) Oct 09 08:46:51 I see no reason but may be hrw or Koen have some :) Oct 09 08:47:52 good morning Oct 09 08:48:22 recalcati_: hi Oct 09 08:48:32 hi mckoan Oct 09 08:48:43 how are you? Oct 09 08:48:52 I want to run lighty with php fastcgi.. I see there is a fastcgi recipe, and a php recipe but now sure how they work with together. Im I on the right track? Ie, are these the packages I need? Oct 09 08:50:22 tasslehoff: I prefer to be in builddir because I often have 2-3 builds running on same machine Oct 09 08:50:30 or is php dependent on the fastcgi package anyway to adding php would be enought? Oct 09 08:50:35 enough Oct 09 08:50:46 recalcati_: fine thx, too busy with ugly bugs making me late Oct 09 08:50:58 hrw, ok Oct 09 08:51:04 yes, I know, it is also my pb Oct 09 09:03:34 oh, Obama got the Nobel peace price Oct 09 09:03:54 That was fast Oct 09 09:12:15 php-5.2.6 doesnt compile. Oct 09 09:25:04 * mckoan_ had some problems with UPS :-) Oct 09 09:34:28 mckoan: UPS hits again? Oct 09 09:38:26 hrw: here there is a summer storm now, BTW in this week I faced to various problems, like Router failure (1 week to replace) too Oct 09 09:38:56 ~curse Telecom operators Oct 09 09:38:57 May the fleas of a thousand camels infest your most sensitive regions, Telecom operators ! Oct 09 09:39:04 mckoan_: shit happens Oct 09 09:47:44 03Frederik Sdun  07shr/import * rad784b442c 10openembedded.git/recipes/navit/navit.inc: Oct 09 09:47:44 Navit: Add gypsy depenciy for shr Oct 09 09:47:44 Signed-off-by: Frederik 'playya' Sdun Oct 09 09:47:44 Signed-off-by: Klaus Kurzmann Oct 09 09:47:45 03Frederik Sdun  07shr/import * r2b48fcd566 10openembedded.git/recipes/e17/exhibit_svn.bb: Oct 09 09:47:47 Use OLD/ SRC_URI for exhibit Oct 09 09:47:49 Signed-off-by: Frederik 'playya' Sdun Oct 09 09:47:51 Signed-off-by: Klaus Kurzmann Oct 09 09:47:53 03Frederik Sdun  07shr/import * r87f9517d49 10openembedded.git/recipes/navit/ (files/use-fso-raw.patch navit.inc): Oct 09 09:47:56 Navit: use fsoraw for Display and GPS Oct 09 09:47:58 Signed-off-by: Frederik 'playya' Sdun Oct 09 09:48:00 Signed-off-by: Klaus Kurzmann Oct 09 09:48:01 did anyone build binutils-cross-sdk_2.18 lately? Oct 09 09:48:06 03Frederik Sdun  07shr/import * rcbaee5a57b 10openembedded.git/recipes/navit/ (files/navit.service navit.inc): Oct 09 09:48:09 Navit: Add dbus activation file Oct 09 09:48:11 Signed-off-by: Frederik 'playya' Sdun Oct 09 09:48:13 Signed-off-by: Klaus Kurzmann Oct 09 09:48:15 03Frederik Sdun  07shr/import * r2a02100840 10openembedded.git/recipes/vala/ (6 files in 2 dirs): Oct 09 09:48:20 Add more patches for vala 0.7.7 Oct 09 09:48:22 Signed-off-by: Frederik 'playya' Sdun Oct 09 09:48:24 Signed-off-by: Klaus Kurzmann Oct 09 09:48:26 03Thomas Zimmermann  07shr/import * rc115412f1b 10openembedded.git/recipes/xorg-xserver/ (2 files in 2 dirs): Oct 09 09:48:29 xserver-xorg: Add patch to rotate touchscreen together with screen (thx to daniels from #xorg-devel) Oct 09 09:48:35 Signed-off-by: Klaus Kurzmann Oct 09 09:48:37 03Frederik Sdun  07shr/import * r399ca5b31b 10openembedded.git/recipes/freesmartphone/libeflvala_git.bb: Oct 09 09:48:40 remove libeflvala recipe from freesmartphone Oct 09 09:48:42 Signed-off-by: Frederik 'playya' Sdun Oct 09 09:48:44 Signed-off-by: Klaus Kurzmann Oct 09 09:48:46 03Frederik Sdun  07shr/import * r2842c31b09 10openembedded.git/recipes/navit/ (files/fr-default-config.patch navit.inc): Oct 09 09:48:53 Navit: Add config file for freerunner Oct 09 09:48:55 Signed-off-by: Frederik 'playya' Sdun Oct 09 09:48:57 Signed-off-by: Klaus Kurzmann Oct 09 09:48:59 03Frederik Sdun  07shr/import * r83d70642d2 10openembedded.git/recipes/python/python-epsilon_svn.bb: Oct 09 09:49:04 Use OLD/ SRC_URI for python-epsilon Oct 09 09:49:06 (2 lines omitted) Oct 09 09:55:11 03Martin Jansa  07shr/import * ra15f352437 10openembedded.git/ (46 files in 6 dirs): Oct 09 09:55:11 Latest xorg stuff 2009-10-07 Oct 09 09:55:11 Signed-off-by: Klaus Kurzmann Oct 09 09:55:12 03Martin Jansa  07shr/import * r24c26c3120 10openembedded.git/ (13 files in 4 dirs): Oct 09 09:55:12 Latest xorg stuff 2009-09-10 Oct 09 09:55:14 Signed-off-by: Klaus Kurzmann Oct 09 09:59:13 I see that shr branch has lot of interesting stuff not present in .dev Oct 09 10:00:25 such is the nature of branches Oct 09 10:01:38 I know Oct 09 10:10:03 how can i solve problems in configure.ac when using something like this: $PKG_CONFIG --variable=vapidir vala-1.0 Oct 09 10:23:00 can I use environment variables in local.conf? Oct 09 10:23:21 export them in my setup-oe.sh, and use them for setting paths in local.conf Oct 09 10:24:25 you can Oct 09 10:24:44 possible that you will need to add them to BB_ENV_EXTRAWHITE Oct 09 10:30:35 hrw, what I want is to set OEBASE="/data/src/oe/" in my script, and then do DL_DIR = "${OE_BASE}/downloads" in local.conf Oct 09 10:30:52 * tasslehoff is spending way too much time on these scripts... Oct 09 10:31:31 tasslehoff: export OEBASE="/data/src/oe/" BB_ENV_EXTRAWHITE="OEBASE" in your script Oct 09 10:32:00 akheron, my humble thanks :) Oct 09 10:37:27 hrw, last script question (I hope). you put OE_DIR in BBPATH in your script. is that necessary? I would like local.conf to specify which oe-tree w/recipes to use. Oct 09 10:38:20 tasslehoff: bitbake uses BBPATH to find local.conf Oct 09 10:38:54 tasslehoff, yes, but OE_DIR points into the openembedded tree. another part of BBPATH points to local.conf Oct 09 10:39:00 hrw, ^^ Oct 09 10:39:13 BBPATH=buildir:oedir Oct 09 10:39:49 bitbake uses BBPATH for: 1. finding config files 2. finding classes Oct 09 10:43:57 hrw, sorry to be slow here, but what do you do when you have built with the dev-branch and what to use stable instead? do you then have to edit and rerun the script (which says OE_DIR=/home/hrw/devel/oe/org.openembedded.dev), and edit local.conf (which says PKGDIR = "/home/hrw/devel/oe/org.openembedded.dev/packages/") to point to your stable oe-tree instead? Oct 09 10:44:21 tasslehoff: my stable/2009 based build has own copy of OE Oct 09 11:02:30 [161267.980000] Alignment trap: Xfbdev (30338) PC=0x40166ed0 Instr=0xe4942004 Address=0x0013b831 FSR 0x011 Oct 09 11:02:34 known? Oct 09 11:03:23 [161348.310000] Alignment trap: X (30354) PC=0x40168ed0 Instr=0xe4942004 Address=0x00249021 FSR 0x011 Oct 09 11:29:30 jo Oct 09 11:29:35 ~seen pb__ Oct 09 11:29:37 pb__ is currently on #oe (24m 48s), last said: 'such is the nature of branches'. Oct 09 11:29:50 ping pb__ Oct 09 11:30:02 hello Oct 09 11:30:20 elo woglinde Oct 09 11:30:42 pb do you remeber how big/small the params partition for bootldr can be? Oct 09 11:30:45 hi hrw Oct 09 11:31:00 minimum is one flash block Oct 09 11:31:12 I guess that's something like 256k for simpad nor Oct 09 11:31:28 okay lower It can bee? Oct 09 11:31:37 hm ah no 128KiB Oct 09 11:31:55 yeah, 128k could be right Oct 09 11:32:11 I think it's 256k on the ipaq because the flash is double-banked, maybe it is only single width on simpad Oct 09 11:34:42 make it 200MB as on Mac Oct 09 12:12:48 Am new to and actually finding OE very good. How can I do a menuconfig and just rebuild the kernel image/modules? Oct 09 12:13:42 cjp: bitbake -c menuconfig virtual/kernel Oct 09 12:14:40 verbatim? Oct 09 12:17:07 koobe: and then to rebuild just that kernel +modules? Oct 09 12:18:34 cjp: perhaps you need to -c clean first, and after menuconfig you might want to copy the resulting .config as defconfig for your board. Oct 09 12:19:29 args Oct 09 12:19:30 I've tried this, however the resultant image always reverts to using the orig config. It never seems to make it into the git directory Oct 09 12:19:46 it dont saves my param Oct 09 12:20:45 I just pulled from dev and cannot compile: recipes/libxml/libxml2-native_2.7.3.bb, do_package) failed Oct 09 12:21:13 cjp: where did you copy your .config? Oct 09 12:21:53 cjp: which board (machine) and kernel is that? Oct 09 12:22:24 cp .config ${OVEROTOP}/user.collection/recipes/linux/linux-omap3-2.6.30/overo/defconfig Oct 09 12:22:24 --- I'm on overo Oct 09 12:22:39 copy from ${OVEROTOP}/tmp/work/overo-angstrom-linux-gnueabi/linux-omap3-2.6.30-r42/git Oct 09 12:23:12 the .config is definately being changed and correct Oct 09 12:23:36 after the copy the build takes place but it reverts to the old config :-( Oct 09 12:26:03 strange things is ... \tmp\work\overo-angstrom-linux-gnueabi\linux-omap3-2.6.30-r47\defconfig is correct BUT \tmp\work\overo-angstrom-linux-gnueabi\linux-omap3-2.6.30-r47\git\.config is the orig version. i.e. without my change Oct 09 12:26:58 I've just looked again, after I followed your suggested 'bitbake -c menuconfig/kernel - it appears to have edited the .config in the \tmp\work\overo-angstrom-linux-gnueabi\linux-omap3-2.6.30-r47\git\ directory. This is expected? Oct 09 12:27:18 Sorry - please ignore last question - dohh Oct 09 12:28:32 yes Oct 09 12:29:38 ping pb Oct 09 12:30:00 hms Oct 09 12:30:12 seems I need the source code for bootldr for simpad again Oct 09 12:32:59 My compile fails in openssl-native_0.9.8j.bb. Is there a website where I can see if the latest git is any good? Oct 09 12:34:36 hi woglinde Oct 09 12:37:34 hm seems I find it Oct 09 12:37:42 http://sourceforge.net/projects/simpad/files/Bootloader/bootldr-simpad01/bootldr-simpad01.tar.bz2/download Oct 09 12:38:17 pb I needed the simpad bootldr code Oct 09 12:38:21 I think I found it Oct 09 12:46:53 righto Oct 09 12:47:11 hm Oct 09 12:47:23 I wonder because my has version 13 instead of 12 Oct 09 12:47:34 but maybee we changed it Oct 09 12:47:50 zecke and me as we workdaorund my bad flash Oct 09 12:48:11 ah Oct 09 12:48:28 or you gave us a private modified version Oct 09 12:48:37 that is possible Oct 09 12:48:49 but I guess you dont have a copy Oct 09 12:48:54 anymore Oct 09 12:49:23 I might do, though most likely on my computer at home Oct 09 12:50:21 the bootldr sources that I have here seem to be version 2.20.2 Oct 09 12:50:46 hm thats older Oct 09 12:50:49 ah, hm, now I found 2.21.14 Oct 09 12:50:53 Compaq OHH BootLoader, Rev 2-21-13 [BIG_KERNEL] [MD5] [MONO] Oct 09 12:50:59 2.21.14: cdm 30-mar-04 Oct 09 12:50:59 -- Update the os_qnx.mk makefile. Oct 09 12:50:59 -- Move ctype.h to be before the internal ctype-is-busted checks. Oct 09 12:50:59 -- Removed the multi-string error for gcc 3.3 Oct 09 12:51:07 2.21.13: pb 03-dec-03 Oct 09 12:51:07 -- use the same value for GAFR0z that the kernel does. Oct 09 12:51:15 do I need a cross compiler? Oct 09 12:51:46 yes, but any arm-linux-gcc should be ok Oct 09 12:52:09 hehe I will see what proting to gcc-4.4 needs Oct 09 12:52:29 looks like 3.3 was the latest we tested with, and jamey seemed to still be on 2.95 judging from the changelog Oct 09 12:52:31 okay .14 is the latest in cvs too Oct 09 12:57:33 hm upstream has now simpad stuff Oct 09 13:12:00 03Marcin Juszkiewicz  07stable/2009 * r0240674224 10openembedded.git/recipes/xserver-common/ (10 files in 2 dirs): (log message trimmed) Oct 09 13:12:00 xserver-common: added 1.30 + patches Oct 09 13:12:00 xserver-common 1.30 is experimental, do not use by default Oct 09 13:12:00 xserver-common 1.30: -screen option is only for Kdrive Oct 09 13:12:00 xserver-common 1.30: another set of updates Oct 09 13:12:02 0024 moves functionality of /etc/X11/Xserver to separate file (except Oct 09 13:12:06 running X11 server) - this allows to use xserver-common package in Oct 09 13:12:50 03Graham Gower  07org.openembedded.dev * r30a06f663f 10openembedded.git/recipes/matchbox-panel/ (2 files in 2 dirs): (log message trimmed) Oct 09 13:12:50 matchbox-panel-0.9.3: fix uninitialised variable causing SIGBUS Oct 09 13:12:50 As reported here: Oct 09 13:12:50 http://lists.o-hand.com/matchbox/0289.html Oct 09 13:12:50 I've not experienced the bug for over a month since applying this change. Oct 09 13:12:52 Signed-off-by: Graham Gower Oct 09 13:12:54 Signed-off-by: Holger Hans Peter Freyther Oct 09 13:18:10 pb could you send me your version too? Oct 09 13:20:04 the sources I have here seem to be unmodified from cvs Oct 09 13:20:13 I'll have a look on my other machines later and see if I have anything else there. Oct 09 13:21:57 mckoan: ping Oct 09 13:22:15 are you going to ELC ? Oct 09 13:22:46 recalcati_: yes Oct 09 13:23:00 Will you go the 15th? Oct 09 13:23:38 03Konrad Mattheis  07org.openembedded.dev * r0fe7e7da98 10openembedded.git/ (3 files in 2 dirs): Oct 09 13:23:38 u-boot 2009.08 Oct 09 13:23:38 Hi, Oct 09 13:23:38 I have this patch to get the new Version of u-boot working. Please could you apply this an the repository or tell me how I can do this. Oct 09 13:23:39 bye Oct 09 13:23:40 pb many thanks Oct 09 13:23:41 Konrad Mattheis Oct 09 13:23:43 I am going on 14th early morning, I've a meeting in the afternoon Oct 09 13:23:43 Signed-off-by: Marcin Juszkiewicz Oct 09 13:23:56 recalcati_: you are going too? Oct 09 13:24:03 I try Oct 09 13:27:57 recalcati_: how long are you staying? Oct 09 13:28:04 one day surely Oct 09 13:28:16 you? Oct 09 13:28:30 recalcati_: so we could meet at WTC Oct 09 13:28:45 recalcati_: I come back on saturday Oct 09 13:29:10 mckoan, recalcati: I will be Tue->Sat Oct 09 13:30:18 recalcati_: as you probably already know, Alessandro will be there too Oct 09 13:30:37 hrw: nice! maybe we will meet Oct 09 13:30:53 mckoan: yes, he told me you are going Oct 09 13:31:42 I try friday, because Rubini FIQ I akready know Oct 09 13:34:39 recalcati_: indeed!, but repetita iuvant :-D Oct 09 13:43:50 http://marcin.juszkiewicz.com.pl/2009/10/09/using-packages-does-not-hurt/ Oct 09 14:14:18 khem: when will the linux->linux-gnu be reverted ? Oct 09 14:23:44 when i do a bitbake -i and type which virtual/kernel i get a list of linux kernels but i dont find a * mark against any of them Oct 09 14:38:39 Anyone else having trouble building the latest from org.openembedded.dev? I get errors in openssl - http://pastebin.ca/1607188 Oct 09 14:44:35 any arm wizard here? I have a compile problem in asm/byteorder.h and gcc 4.3.3. Oct 09 14:44:47 offending line: asm ("eor\t%0, %1, %1, ror #16" : "=r" (t) : "r" (x)); Oct 09 14:44:59 error: expected ')' before ':' token Oct 09 14:45:06 effem ah Oct 09 14:45:14 I hit this too some times Oct 09 14:45:22 google there is a patch for Oct 09 14:45:33 woglinde: cool, thanks Oct 09 14:45:39 ah If I remeber Oct 09 14:45:45 it was badly stlye Oct 09 14:45:55 and forbidden writing asm code this way Oct 09 14:48:48 03Martin Jansa  07org.openembedded.dev * r117ec4def4 10openembedded.git/ (188 files in 14 dirs): Oct 09 14:48:48 Add latest xorg stuff for xserver-1.7, script for easy recipe bump and xf86-video-glamo-kms Oct 09 14:48:48 Signed-off-by: Marcin Juszkiewicz Oct 09 14:48:49 03Marcin Juszkiewicz  07org.openembedded.dev * r28dd4f870c 10openembedded.git/conf/checksums.ini: checksums.ini: some new entries Oct 09 14:48:58 o fsck. Oct 09 14:49:16 I did not plan to push Martin's one Oct 09 14:49:47 03Marcin Juszkiewicz  07org.openembedded.dev * rf39bf1b386 10openembedded.git/ (188 files in 14 dirs): Oct 09 14:49:47 Revert "Add latest xorg stuff for xserver-1.7, script for easy recipe bump and xf86-video-glamo-kms" Oct 09 14:49:47 This reverts commit 117ec4def4dff7abad394279698e1ddaef9407a7. Oct 09 14:55:06 mckoan: ping Oct 09 14:57:42 morning Oct 09 14:57:51 yo kergoth Oct 09 14:59:30 hi Oct 09 14:59:59 hrw: hi, I was surprised too, that you pushed it :) Oct 09 15:00:30 hrw: It seems that 15th is better for openebedded and android, isn't it? Oct 09 15:00:46 hm Oct 09 15:00:49 recalcati: 15th is definitelly better as I will talk then Oct 09 15:00:53 hrw: what about libdrm_git.bb as only one provider, only issue would be that we have to use anothere sane-srcrev for om-gta02 machine, because its from another git repository Oct 09 15:00:54 yo kergoth Oct 09 15:01:11 pb are you familar with Error: misaligned branch destination? Oct 09 15:01:24 woglinde: not intimately, but I can guess what it means Oct 09 15:01:35 something with labels Oct 09 15:01:45 JaMa: I am fine with two recipes - you need to provide PREF_PROV for repos or in other way make glamo version not default Oct 09 15:02:02 boot-sa1100.s has Oct 09 15:02:04 DataAbortEntryPoint: b HandleDataAbort Oct 09 15:02:30 right, and what does the code around HandleDataAbort look like? Oct 09 15:02:33 where HandleDataAbort is definied later in the file Oct 09 15:03:00 hrw: but maybe it would be easier to switch back to upstream libdrm for om-gta02 as soon as Weiss's patches for glamo gets upstream Oct 09 15:03:26 pb -> http://paste.debian.net/48620/ Oct 09 15:03:41 needs it a simple .align 2 in the front? Oct 09 15:03:42 woglinde: what does the code before that look like? Oct 09 15:04:05 pb hm? Oct 09 15:04:26 the code prior to "HandleDataAbort" in the file Oct 09 15:04:38 i.e. what are the lines before the first one that you pasted Oct 09 15:05:04 there is no one which calls it Oct 09 15:05:08 only the entry point Oct 09 15:05:32 so you are saying that "HandleDataAbort:" is the very first line in the source file? Oct 09 15:05:35 is the output after cpp usefull? Oct 09 15:05:45 woglinde can't find patch Oct 09 15:05:53 hm maybee I am not correct with asm-files Oct 09 15:05:57 no idea waht is wrong Oct 09 15:06:08 I thought it is like as c functions Oct 09 15:06:15 ah okay Oct 09 15:06:21 sorry I posted it above Oct 09 15:06:23 no, looked at some different files, from differnet sources and they all have this asm line Oct 09 15:06:28 will do on pastebin again Oct 09 15:07:26 pb -> http://paste.debian.net/48621/ Oct 09 15:07:38 effem cost me some time to find it Oct 09 15:07:51 effem could you pastebin it? Oct 09 15:07:51 ok, np Oct 09 15:07:55 woglinde: so, those are the lines immediately prior to "HandleDataAbort:" in the file? Oct 09 15:07:59 sure Oct 09 15:08:06 pb yes Oct 09 15:08:13 only occurs this twice Oct 09 15:08:24 I'm not sure you understand what I am saying. Oct 09 15:08:34 forget about the calling order. I am not interested in what branches to what. Oct 09 15:08:43 hm Oct 09 15:08:44 sorry Oct 09 15:09:01 What I want you to show me is the actual contents of the file. In other words, if you put your cursor on the line of the source file that has "HandleDataAbort:", and then press the cursor up key, what line are you on now? Oct 09 15:09:59 hm ah Oct 09 15:10:02 there is tab Oct 09 15:10:08 maybee this is the problem? Oct 09 15:10:13 tabs shouldn't matter Oct 09 15:10:21 can you put the preceding ten or so lines in a pastebin? Oct 09 15:10:22 woglinde: http://www.pastebin.ca/1607259 Oct 09 15:11:05 yes Oct 09 15:11:29 http://paste.debian.net/48623/ Oct 09 15:12:10 eFfeM and the code as it is written in the file Oct 09 15:12:27 in the .h file ? Oct 09 15:12:48 the offending line is in the end but I can pastebin the whole .h file (or did you want .c) Oct 09 15:13:00 woglinde: that looks ok. how about the next block of code up? Oct 09 15:13:08 pb more above? Oct 09 15:13:12 right Oct 09 15:13:17 before those .words Oct 09 15:13:21 hm I will paste whole file Oct 09 15:13:26 ok, good idea Oct 09 15:13:58 http://paste.debian.net/48624/ Oct 09 15:15:11 woglinde: this has the full byteorder.h Oct 09 15:15:11 hm Oct 09 15:15:12 http://www.pastebin.ca/1607273 Oct 09 15:15:27 that all looks basically ok Oct 09 15:15:29 for now compiled the file with -D__thumb__ to get rid of the code Oct 09 15:15:34 you haven't changed BL_SIG or anything, have you? Oct 09 15:15:46 pb I havent changed something Oct 09 15:15:56 only downloaded from the side Oct 09 15:16:11 and make sure It finds linkage.h Oct 09 15:22:46 effem I think the \t is wrong Oct 09 15:23:00 ah, ok will try to replace that Oct 09 15:24:18 but I am not sure Oct 09 15:25:18 jepp Oct 09 15:25:51 woglinde: replacing \t by space did yield the same error Oct 09 15:32:15 effem -> http://209.85.129.132/search?q=cache:ewSHQlmYEs0J:www.ethernut.de/en/documents/arm-inline-asm.html+asm+"ror+%2316"&cd=2&hl=de&ct=clnk&client=iceweasel-a Oct 09 15:33:24 effem rewrite according to it Oct 09 15:33:26 -> http://pastebin.ca/1607313 Oct 09 15:36:18 pb any idea? Oct 09 15:36:51 no, not at the moment. the code looks ok to me. Oct 09 15:37:02 might be a binutils bug, or might be some subtle thing in the code upsetting it that isn't obvious at a first inspection. Oct 09 15:37:26 it did use to work, obviously, so if you haven't modified anything from cvs I would incline towards suspecting that the assembler is broke Oct 09 15:37:27 hm is code after cpp usefull? Oct 09 15:37:44 pb its the latest from oe Oct 09 15:37:46 yes, that might be worth a look Oct 09 15:38:07 yah, latest from oe is obviously quite a lot newer than what bootldr was tested with Oct 09 15:38:12 pb -> http://paste.debian.net/48630/ Oct 09 15:38:15 pb I know Oct 09 15:38:24 but I am to lame to compile gcc-3.3 yet Oct 09 15:38:32 but maybee I hat been faster Oct 09 15:38:33 ah Oct 09 15:38:39 kernel macro? Oct 09 15:38:45 .ascii "SIMPAD" Oct 09 15:38:46 that is no good Oct 09 15:38:48 that's BL_SIG Oct 09 15:39:09 thats need a align? Oct 09 15:39:15 where did that come from? Oct 09 15:39:19 it isn't in my cvs copy Oct 09 15:39:28 thats the simpad copy Oct 09 15:39:33 aeh reworking Oct 09 15:39:47 oh, you are using simpad cvs not handhelds.org cvs? Oct 09 15:40:01 from here -> http://sourceforge.net/projects/simpad/files/Bootloader/bootldr-simpad01/bootldr-simpad01.tar.bz2/download Oct 09 15:40:04 well, anyway, that definition for BL_SIG is no good. it must be precisely four characters. Oct 09 15:40:10 hm I thought it was different for simpad Oct 09 15:40:26 or can I use it Oct 09 15:40:38 there are some changes that's look worth to me Oct 09 15:40:41 dunno, I don't really remember anymore Oct 09 15:40:46 hm Oct 09 15:41:18 last question can I boot a bootldr from bootldr? Oct 09 15:41:36 in theory yes, in practice that never used to work very well for me Oct 09 15:42:05 because of init stuff right? Oct 09 15:43:08 hm this is definied in bootflags Oct 09 15:43:21 http://paste.debian.net/48632/ Oct 09 15:43:33 hms Oct 09 15:44:13 http://paste.debian.net/48634/ Oct 09 15:44:53 that's fine, it is only BL_SIG that is wrong Oct 09 15:45:23 args sorry Oct 09 15:45:25 silly me Oct 09 15:52:36 jo robert Oct 09 15:55:46 hms Oct 09 15:55:53 need to rewrite this Oct 09 15:55:54 temp = *(++((u32 *) in)); Oct 09 15:56:56 ah okay Oct 09 15:56:58 with +1 Oct 09 16:08:55 woglinde thanks, will try Oct 09 16:25:20 woglinde: tried your suggestion but the multiline split did not make a difference. It now barfs at the line with the first : (the one after asm). Oct 09 16:25:36 compiling with -D__thumb__ works as a workaround for me Oct 09 16:28:23 bye all Oct 09 16:35:32 03Jeremy Lainé  07org.openembedded.dev * r924dc8e21a 10openembedded.git/recipes/linux/ (4 files in 2 dirs): linux-2.6.29: update boc01 isl12024 driver to new style i2c Oct 09 16:35:39 effem hm Oct 09 16:35:45 whats the error message now? Oct 09 16:35:49 exactly Oct 09 16:38:17 woglinde: Oct 09 16:38:18 /usr/include/asm/byteorder.h: In function '___arch__swab32': Oct 09 16:38:19 /usr/include/asm/byteorder.h:34: error: expected ')' before ':' token Oct 09 16:39:37 line 32 is the old asm line commented out; line 33-37 is the asm line as you gave it Oct 09 16:41:54 woglinde: guess it is specific for this gcc, will move to something newer anyway Oct 09 16:42:11 to mne the line seems in line in what the cookbook says Oct 09 16:42:34 effem hm could you remove the \t\n Oct 09 16:42:42 I think its not nessacry Oct 09 16:42:48 in oneliner Oct 09 16:43:19 hm and I changed the order Oct 09 16:59:37 woglinde: remove \t\n did not make a difference Oct 09 16:59:43 hms Oct 09 16:59:53 same error Oct 09 17:05:03 anyway, calling it a day, cya all tomorrow! Oct 09 17:05:57 args Oct 09 17:06:06 I should have read right Oct 09 17:06:22 __asm__ volatile is needed Oct 09 17:06:25 hms Oct 09 17:06:57 args nope Oct 09 17:07:01 __asm_ only Oct 09 17:07:04 okay Oct 09 17:17:23 morning all Oct 09 17:18:08 03Jeremy Lainé  07org.openembedded.dev * r8b7c982303 10openembedded.git/recipes/linux/linux-2.6.29/boc01/005-091008-isl12024.patch: linux-2.6.29: update boc01 unique ID patch to use sysfs Oct 09 17:26:42 he thesing Oct 09 17:26:52 thesing eabi userland works Oct 09 17:27:09 but we have to set -W,-fixthearmv4 Oct 09 17:27:12 thats great Oct 09 17:27:15 as compiler flag for now Oct 09 17:27:30 because gcc missing this in his spec Oct 09 17:27:37 was it already changed? Oct 09 17:29:36 khem is working on a patch Oct 09 17:29:55 and I didnt check something in oe for now Oct 09 17:37:56 ok Oct 09 18:15:08 jo kergoth Oct 09 18:15:21 hey Oct 09 18:38:08 pb_ and others, I have sent 3 patches which revert the linux to linux-gnu conversion. have a look at them I tested them to best I could by building for mipsel arm and powerpc and minimal and angstrom distros Oct 09 18:44:17 ok, thanks Oct 09 18:44:26 it would be nice to have at least one architecture tested with micro as well Oct 09 19:05:25 ok let my current build finish then I will fire micro/osk5912/arm/gnueabi Oct 09 19:38:07 woglinde: around ? Oct 09 19:38:59 woglinde: Here is untested patch which should pass the option to linker I would appreciate any help in testing it out http://sakrah.homelinux.org/Public/gcc-armv4-pass-fix-v4bx-to-ld.patch Oct 09 19:53:44 he khem Oct 09 19:55:04 hehe small fix Oct 09 20:08:18 yes try it out rebuild gcc-cross and check the command cross-gcc -march=armv4 -v -x c /dev/null Oct 09 20:08:55 and then inspect the output especially look at what is passed on commandline to collect2 Oct 09 20:09:45 maybe tomorrow Oct 09 20:10:11 no worried if I find time I will try to do it Oct 09 20:10:16 my day is still half Oct 09 20:21:53 Bah Oct 09 20:22:30 Anyone had HOST_OS not get munged to linux-gnu, on x86_64 ? Oct 09 20:26:17 khem? Oct 09 20:36:31 Tartarus: yah, I think that was one of the outstanding problems with the linux-gnu thing. khem posted a patch to revert those changes and go back to just "linux" which will probably clear it up. Oct 09 20:43:01 Tartarus: actually it worked all ok it was one liner to fix it but I decided to revert it Oct 09 20:43:19 the patches are already on list I am testing various builds now Oct 09 20:43:24 so far all look ok Oct 09 20:45:04 its the way we grab OS value in bitbake.conf BUILD_OS = "${@os.uname()[0].lower()}" Oct 09 20:45:25 will set it to linux and not to linux-gnu Oct 09 21:55:27 hmm, do we have a build-essential equivalent? Oct 09 21:58:55 kergoth I fear not Oct 09 21:59:07 if that's the case, we should fix that Oct 09 22:00:36 I am running on an avr32 with uclibc - I was having an issue where on boot, the system could not create the /proc filesystem Oct 09 22:00:46 it was complaining that libintl.so was not found Oct 09 22:00:57 i built a console-image Oct 09 22:01:11 I searched the image and it was not part of it Oct 09 22:01:29 after some googling, I found a reference to gettext, so I added that to the image Oct 09 22:01:36 hm Oct 09 22:01:46 now libintl.so is part of the image but when it would boot, it would still not find it Oct 09 22:01:47 which distribution? Oct 09 22:01:56 angstrom 2009.x Oct 09 22:02:37 i ran ldconfig and rebuilt the library path and it was happy about that Oct 09 22:02:50 hm Oct 09 22:02:57 now, however, it hangs on boot Oct 09 22:03:06 init? Oct 09 22:03:10 udev times out Oct 09 22:03:18 try mdev Oct 09 22:03:24 then after that, it hangs here: Oct 09 22:03:25 hwclock: can't open '/dev/misc/rtc': No such file or directory date: invalid date '%2m%2d%2H%2M2009' hwclock: can't open '/dev/misc/rtc': No such file or directory Oct 09 22:03:25 hi bluelightning Oct 09 22:03:38 hi woglinde Oct 09 22:03:46 martinmeba disable hwlock setting for now Oct 09 22:03:51 ok Oct 09 22:03:54 that is what I was thinking Oct 09 22:04:04 woglinde: sorry haven't had a chance to do any work on the opie recipes yet, will try to do something this weekend though Oct 09 22:04:14 bluelightning no stress Oct 09 22:06:12 any idea why the library doesn't make it into the rootfs? Oct 09 22:07:38 which image did you build? Oct 09 22:07:46 console-image Oct 09 22:08:06 hm okay Oct 09 22:08:16 seems no app has a reald dep on it Oct 09 22:08:25 i'm trying to take baby steps =D Oct 09 22:08:30 so some app is missing a RDEP Oct 09 22:09:43 bluelightning: does mplayer work with opie? Oct 09 22:09:58 thesing: it used to yes Oct 09 22:10:05 I haven't tried it recently Oct 09 22:10:17 and is there a gui for it? Oct 09 22:10:18 there's even a QtE frontend for it somewhere around Oct 09 22:10:24 ^ :) Oct 09 22:10:29 this kino2? Oct 09 22:10:34 yeah that's the one Oct 09 22:10:56 marex and I were trying to get it fixed up at one point, can't recall where we got to Oct 09 22:11:06 the webpage is dead and I didn't find the sourcecode somewhere else. Oct 09 22:11:22 I think I may have the latest available source somewhere Oct 09 22:11:31 thesing opie is in git now Oct 09 22:11:54 bluelightning: be sure to don't lose it. Oct 09 22:12:09 http://gitorious.org/opie/opie/commits/master Oct 09 22:12:11 thesing: well, if it's not here then it's back on my PC in NZ Oct 09 22:12:30 bluelightning: write it to multiple CDs and send to friends ;) Oct 09 22:12:42 thesing: which means it may not immediately be to hand, but I can still get access to it :) Oct 09 22:12:44 thesing I am making my work through bootldr *g* Oct 09 22:13:03 woglinde: woah, working on bootldr? the compaq one? Oct 09 22:13:11 surely not Oct 09 22:13:20 bluelightning yes Oct 09 22:13:29 args Oct 09 22:13:32 I meant Oct 09 22:13:39 I am working through Oct 09 22:13:40 args Oct 09 22:13:42 woglinde: did that bootloader pass the params at the 'old' 2.4 location? Oct 09 22:13:52 ah ok Oct 09 22:14:05 ant hm dont know but 2.6 kernels boots Oct 09 22:14:08 or at the new one 'atag-compatible'? Oct 09 22:14:12 ah, ok Oct 09 22:14:17 but the code is ugly Oct 09 22:14:22 in some files Oct 09 22:14:24 and 2.4 recognizes cmdline? Oct 09 22:14:27 mixing to much ob Oct 09 22:14:29 upo Oct 09 22:14:38 bootloaders code is always ugly :) Oct 09 22:14:44 ant kernel or bootloader? Oct 09 22:14:50 bootloader Oct 09 22:14:50 jay7 hm I mean style Oct 09 22:14:54 not function Oct 09 22:15:04 kernel expects them in a new location since couple of yrs Oct 09 22:15:07 woglinde: I mean style too :) Oct 09 22:15:15 woglinde: why do you do this? Oct 09 22:15:46 thesing because zecke and I patched the bdlr once to workaround some flash errors Oct 09 22:16:01 and now I cann't make a params partition Oct 09 22:17:48 so I am trying to build another bootldr Oct 09 22:43:01 woglinde: I tested my gcc patch and it works !! Oct 09 22:43:40 woglinde: tried with eabi eabi+armv4 oabi combos and everytime it passes the right options to linker now Oct 09 22:44:05 argh Oct 09 22:44:08 mickey|sports: ping Oct 09 22:44:17 (or anyone else famiilar with the python recipes/build) Oct 09 22:44:40 kergoth: we can try :) Oct 09 22:44:58 build=i686 target=x86_64. kaboom. Oct 09 22:45:28 got it building, but it fails to build what may be all of the c extensions.. Oct 09 22:45:29 hmm Oct 09 22:45:41 it doesn't want to import the 64 bit objects Oct 09 22:45:58 maybe the python-native needs to return the target arch/os rather than build, or something.. Oct 09 22:46:02 * kergoth pokes at sysconfig Oct 09 22:49:12 kergoth is host=x86 or x86_64 Oct 09 22:49:54 i just told you, the build machine is i686, the target is x86_64. that tells you that python-native's HOST_ARCH is i686, python's is x86_64 Oct 09 22:50:51 make sure that is what is gets Oct 09 22:51:38 so its the target recipe thats troubling right ? Oct 09 22:52:57 yes, but it relies on the native one to supply things, some of which have to do with the target Oct 09 22:53:29 03Andrea Adami  07org.openembedded.dev * r1aadd3a3f4 10openembedded.git/recipes/kexecboot/kexecboot_git.bb: kexecboot_git: move to e7f85b13e144a594fa726f5f7ca4bd3e2327e2b3. Bump PR. Oct 09 22:53:29 during build ? Oct 09 22:53:32 it asks the native python's distutils for the library path, and it's patched to return the current STAGING_LIBDIR, and replaces BUILD_SYS with HOST_SYS in places to match the target Oct 09 22:53:35 yep Oct 09 22:53:39 03Andrea Adami  07org.openembedded.dev * r65fa06e7ec 10openembedded.git/recipes/kexecboot/linux-kexecboot.inc: linux-kexecboot.inc: to be quiet we set master console on serial. Bump PR. Oct 09 22:58:35 kergoth: hmmm I think the issue could be 32bit host and 64bit target Oct 09 22:59:11 yep, thats what i was thinking, i dunno. it seems weird that it can build for a compeltelyd ifferent arch fine, but not i686->x86_64? makes no sense Oct 09 22:59:13 hrmph Oct 09 22:59:14 python may not be as robust as bfd when it comes to cross compiling and cross hosting Oct 09 22:59:18 urg, why does omap3 take 250nS between reading pairs of words from gpmc .... Oct 09 23:00:15 good nite Oct 09 23:02:12 kergoth: is distutils compiled for target ? iow is it cross compiled ? Oct 09 23:02:31 kergoth: I think you need to cross compile distutils like gcc etc. Oct 09 23:02:32 * kergoth doesn't know that yet, not very familiar with the build Oct 09 23:04:43 kergoth: it seems to me not Oct 09 23:05:04 there should be some option like enable-targets=all Oct 09 23:05:10 well, why does it work for arm and the like then? Oct 09 23:05:44 remains to be found but prolly its doing the cross arch parts well Oct 09 23:06:13 like I have x86_64 build and 32-bit arm target it works ok Oct 09 23:06:18 this is cross arch, i686 is not x86_64 :) Oct 09 23:06:30 wonder if python has special handling of multilib stuff Oct 09 23:06:34 maybe that could be it Oct 09 23:06:56 yeah what I mean it if its host arch is 32bit then it may not be able to do the 64bit target requests Oct 09 23:06:59 looks like a bunch of the extensions were built for i686 and installed into /usr/lib/python2.6, but most of the stuff went into /usr/lib64/python2.6 Oct 09 23:07:04 night all Oct 09 23:07:37 * kergoth sighs, going to have to actually learn how this build process works.. Oct 09 23:08:41 does it use --plat-name somewhere in configure or build ? Oct 09 23:10:38 is python broken? Oct 09 23:10:59 well, for x86_64 target, for me, anyway. /me tries roman's amd64 branch Oct 09 23:11:41 bother, something about python is failing, I need to investigate ..... Oct 09 23:13:36 kergoth: if you have a x86_64 box you could try that for build machine and see if that works Oct 09 23:13:54 hmm, good idea Oct 09 23:14:22 They did some work on python to get it cross compiling modules atleast on 2.6 Oct 09 23:14:33 but it was on winnt platform Oct 09 23:14:55 one could use 32bit win platform to generate modules that would run on 64bit winnt Oct 09 23:23:18 kergoth: I think this patch is interesting http://dev.gentoo.org/~kumba/tmp/python-2.5.4-cross-compile-alpha5.patch Oct 09 23:24:03 hmm, interesting indeed Oct 09 23:24:07 thanks for the pointer Oct 09 23:32:36 khem: it looks like it compiled everything fine, but mickeyl's python manifest generator doesn't distinguish between platform specific and non-platform specific modules, but the python build does, so some modules are in /usr/lib64/python2.6 and others in /usr/lib/python2.6 Oct 09 23:32:42 * kergoth heads home, will play with it there Oct 09 23:32:58 kergoth: ah Oct 09 23:33:17 (the non platform specific modules end up not packaged as a result, so it seems) Oct 09 23:33:18 * kergoth wanders off Oct 09 23:37:26 python extension modules bbs should inherit distutils always Oct 09 23:45:01 kergoth: it should not install anything under /usr/lib/python2.6 for x86_64 target if it is doing so means the module is not built correctly Oct 09 23:45:29 on multilib machines /usr/lib/python2.6 and /usr/lib64/python2.6 have same modules Oct 09 23:46:15 * khem will try a maiden amd64 build sometime over weekend Oct 09 23:47:34 * khem drives off Oct 10 00:07:17 i'm running into this error that it seems a number of people have run into: http://pastebin.ca/raw/1604782 Oct 10 00:07:47 anyone know what is causing it? I don't see anything obvious even with bitbake -v Oct 10 01:29:59 what is the solution for the key error messages? **** ENDING LOGGING AT Sat Oct 10 02:59:56 2009