**** BEGIN LOGGING AT Wed Dec 12 02:59:58 2012 Dec 12 08:02:40 Hi all! Does anyone no, how I will setup an recovery image (seperate kernel/minimal filesystem) with bitbake? I'd like to setup a recipes for generating the recovery image. Result of the recipe should be the original-build, that is already working plus a revovery image with a modified kernel/fs. My problem is how to integrate this in bitbake. Any ideas about that? Dec 12 08:25:58 good morning Dec 12 09:20:57 Trying to try danny, and get: Missing or unbuildable dependency chain was: ['packagegroup-cross-canadian-arm', 'gcc-cross-canadian-arm', 'gmp-nativesdk'] Dec 12 09:21:02 any suggestions for a cure Dec 12 09:21:05 ? Dec 12 09:38:08 tasslehoff: you are building meta-toolchain I assume? Dec 12 09:38:32 tasslehoff: is there any other error indicating why gmp-nativesdk isn't available? it looks like it should be... Dec 12 09:38:52 morning all Dec 12 09:39:58 bluelightning: http://pastebin.com/nSiMkB7U Dec 12 09:42:08 tasslehoff: which MACHINE ? Dec 12 09:45:22 mckoan: my own machine, but basically a beagleboard. Dec 12 09:51:14 tasslehoff: do you need to be using toolchain-layer ? Dec 12 09:52:17 also, it looks like the distro config expects a linaro toolchain; not sure if there are other settings apart from the PREFERRED_VERSION values that might cause trouble given that it's not there... Dec 12 09:54:53 bluelightning: I get the same when compiling console-image.bb. I use the angstrom setup-scripts, so I'll bug koen some more. Dec 12 09:55:23 tasslehoff: sounds like a good idea in that case Dec 12 10:52:03 I can't access to the login building default core-image-sato. Any issue related to this ? OTOH with core-image-minimal works. Dec 12 10:53:32 eballetbo: you'll need to give more details than that Dec 12 10:59:29 rburton: Starting the core-image-sato image in my board I can't access with root account. Dec 12 11:01:06 did you set the debug-tweaks distro feature, so there is no root password? Dec 12 11:03:12 yes I've debug-tweaks in my local.conf. Dec 12 11:04:19 no problems that i'm aware of Dec 12 11:07:28 ok, thanks. I just updated to the latest git version from today, and I'm rebuilding the core-image-sato. Forget for a while the question, I just wanted to know if someone had the same problem. Probably is an issue with my local tree. Dec 12 11:59:04 rburton: I've the issue with current mainline, any idea where I can look ? Dec 12 12:25:33 compiling both 32-bit and 64-bit sdks from a 64-bit machine, should that work in oe-core? Dec 12 12:26:07 yes Dec 12 12:26:31 SDKMACHINE = "x86_64" Dec 12 12:26:32 #SDKMACHINE = "i686" Dec 12 12:31:38 do i need to pass the sysroot gcc arguments to compile anything at all? Dec 12 12:31:56 hm, yes i think so Dec 12 12:32:05 * rburton wonders where this upstream broke Dec 12 12:36:36 anyone know anything about cmake, and why a build would be ignoring the toolchain.cmake file? Dec 12 12:57:11 hi, all! Dec 12 12:57:46 how I can configure oe-core to include buildbot build number into version, if build is done from buildbot? Dec 12 13:20:04 urg Dec 12 13:20:07 where wa I Dec 12 13:20:40 my image recipe wants a LICENSE and LICENSE checksum Dec 12 13:20:43 when did this start? Dec 12 13:21:38 since oe-core I think Dec 12 13:21:38 Crofton|work: LIC_FILES_CHKSUM shouldn't be needed unless you have somehow re-enabled do_configure Dec 12 13:21:51 for LICENSE just specify MIT Dec 12 13:22:13 yeah, I did the MIT thing, then it bitched about the checksum Dec 12 13:22:40 Tasks Dec 12 13:22:41 ERROR: e200-test-image: Recipe file does not have license file information (LIC_FILES_CHKSUM) Dec 12 13:22:41 ERROR: Licensing Error: LIC_FILES_CHKSUM does not match, please fix Dec 12 13:23:33 how would I have re-enable configure? Dec 12 13:23:58 maybe cleansstate Dec 12 13:24:06 no Dec 12 13:24:27 which version/branch of OE are you using? Dec 12 13:24:34 so the image cless disable configure, but the qa test runs anyway Dec 12 13:24:41 master Dec 12 13:25:02 + meta-oe Dec 12 13:25:31 is it do_configure where it's failing? Dec 12 13:25:54 yep Dec 12 13:26:42 that's not normal for a recipe that inherits image.bbclass... Dec 12 13:27:02 it inherita core-image Dec 12 13:27:09 right, which inherits image Dec 12 13:28:54 Crofton|work: does your image do anything unusual with tasks? Dec 12 13:29:06 no, just a list of packages Dec 12 13:30:19 RP: any ideas why Crofton|work's image would now be running do_configure which is marked noexec? Dec 12 13:30:34 Crofton|work: bitbake master as well I assume? Dec 12 13:30:42 should be Dec 12 13:31:12 weird, there is a local diff Dec 12 13:31:33 which is weird, because I never edit that Dec 12 13:31:35 what is it? Dec 12 13:31:54 looks like bitbake.conf was there and disappeaed Dec 12 13:31:59 or vice versa Dec 12 13:32:09 let me clean it up Dec 12 13:32:26 # deleted: conf/bitbake.conf Dec 12 13:32:57 I cheked out the file and same issue Dec 12 13:33:29 which bitbake.conf though? the one in the bitbake source tree? Dec 12 13:33:35 right Dec 12 13:33:42 that's not used with OE anyway Dec 12 13:35:38 Crofton|work: The search paths for files has changed recently. I'm going to guess that it can't actually find the file anymore with the FILESPATH changes Dec 12 13:36:58 RP: but do_configure shouldn't even be running for images, thus that check should not be running... Dec 12 13:39:06 bluelightning: true. I don't know how it would manage that :/ Dec 12 13:39:25 I started without a LICENSE, failed added one Dec 12 13:39:41 then checksum fail Dec 12 13:39:44 should I nuke everything and start over? Dec 12 13:39:49 no Dec 12 13:40:03 Crofton|work: no, I'd like to understand this Dec 12 13:40:21 The license not set test is in base.bbclass Dec 12 13:40:26 so that is straight forward Dec 12 13:43:21 well, something is foul Dec 12 13:47:19 Crofton|work: I just tested here and it doesn't run configure for images Dec 12 13:48:39 Crofton|work: something else has to be wrong, like its finding the wrong class or not inheriting at all Dec 12 13:48:53 is it getting anything from meta-oe? Dec 12 13:50:06 Crofton|work: can you pastebin bitbake xxx-image -e | grep ^_task_deps= Dec 12 13:51:05 nothing Dec 12 13:51:23 urg Dec 12 13:52:07 http://pastebin.com/nukNBrpB Dec 12 13:52:51 Crofton|work: 'noexec': {'do_package_write': '1', 'do_build': '1'} Dec 12 13:52:59 there are several things missing from there :/ Dec 12 13:53:09 you speak in tongues Dec 12 13:53:17 'noexec': {'do_install': '1', 'do_configure': '1', 'do_package': '1', 'do_package_write': '1', 'do_build': '1', 'do_package_write_ipk': '1', 'do_package_write_deb': '1', 'do_fetch': '1', 'do_populate_sysroot': '1', 'do_patch': '1', 'do_compile': '1', 'do_package_write_rpm': '1', 'do_unpack': '1'} Dec 12 13:53:20 is what I see here Dec 12 13:53:30 ok Dec 12 13:55:58 Crofton|work: could you share the output of bitbake xxx-image | grep ^BBINCLUDED Dec 12 13:56:27 Crofton|work: just trying to track down what exactly its doing (rather than what we think it should be doing) Dec 12 13:57:44 http://pastebin.com/hybaMDwF Dec 12 13:58:39 Crofton|work: It never includes core-image.bbclass or image.bbclass ? Dec 12 13:59:50 rofl Dec 12 13:59:56 oh. you will like this Dec 12 14:00:13 so the include core=image or whatever was the last line Dec 12 14:00:25 and the line before was # tools-debug \ Dec 12 14:00:41 which I had to comment out of image install for some reason Dec 12 14:00:56 so it treated the inclde as a comment Dec 12 14:01:44 is this the "correct" behavior Dec 12 14:02:23 anyway, thanks for the help Dec 12 14:02:30 this was driving me aroun the bend Dec 12 14:03:53 Crofton|work: pretty sure that's a bug... would you mind reporting it? Dec 12 14:04:12 oh god, you mean use bugzilla? Dec 12 14:04:24 I do yes :) Dec 12 14:04:32 I was afraid of that Dec 12 14:09:04 bluelightning: its not bug Dec 12 14:09:22 oh? Dec 12 14:09:40 bluelightning: well, I guess it should probably error if the following line isn't a comment too Dec 12 14:09:41 \ extends comments to the next line? Dec 12 14:09:51 so yes, please file it Dec 12 14:09:55 It does Dec 12 14:10:04 hmm, I wouldn't have expected that behaviour myself Dec 12 14:10:38 to my mind, a comment ought to completely disable anything that follows it Dec 12 14:11:28 maybe we needed continuation through comments when we semi-supported commented-out lines in the middle of a multi-line continuation, but we don't allow that anymore Dec 12 14:12:50 https://bugzilla.yoctoproject.org/show_bug.cgi?id=3570 Dec 12 14:12:58 discuss Dec 12 14:13:44 the behavior will certainly catch others, but from now on, I'll alaways have a blank line after comments :) Dec 12 14:13:44 rofl Dec 12 14:13:52 poky.bs.watcher@.... Dec 12 14:24:45 Crofton|work: it's the bachelor of science watcher ;) Dec 12 14:25:50 that is not what I was thinking :) Dec 12 14:36:39 hey guys. Is there a howto floating around on how to integrate a device into an existing kernel-tree? I'm trying to get support for an archos gen9 tablet working, which is pretty similar to the omap blaze tablet Dec 12 14:38:24 the vendor-kernel is based on a 3.0.21 Android ICS with TI modifications; I tried to fumble it into a 3.5 TI kernel yesterday by just importing the files, but there have been some major changes to the whole omap-support system since then Dec 12 14:44:53 dFence: I suspect it won't be straightforward Dec 12 14:47:56 so far, nothing has been ;) but the vendor-kernel has some major quirks - for once, the MUSB-Stack isn't working as it should, and I'd like to integrate it into OE Dec 12 16:25:27 hi, does anybody know of any glib-2.0_2.32 or greater recipe? Dec 12 16:37:50 ggonzalez: openembedded-core/meta/recipes-core/glib-2.0/glib-2.0_2.34.3.bb Dec 12 16:39:36 wow cool thanks Dec 12 16:40:00 I hope I can run it straight in the over branch Dec 12 16:55:23 khem: ping Dec 12 16:56:27 khem: can you take a look at two patches I sent for toolchain-layer? Dec 12 17:38:26 moin Dec 12 18:21:09 I guess I used wrong subject for bitbake patch Dec 12 18:21:53 I see it in oe patchwork, not bitbake :\ Dec 12 18:23:00 is it ok to use [meta-oe][bitbake-devel] ? Dec 12 18:26:34 wrong ML for start Dec 12 18:26:57 bitbake-devel@lists.openembedded.org Dec 12 18:27:02 whoops Dec 12 18:28:05 the only mail that should ahve [meta-oe] is stuff going into the meta-openembedded repo Dec 12 18:29:45 okkie Dec 12 18:30:28 * eren brb Dec 12 18:40:19 I guess the subject prefix "[PATCH]" is enough for bitbake-devel, mailman automatically adds [bitbake-devel] Dec 12 18:41:17 eren: yes, and thanks for contrib/vim changes, appreciated Dec 12 18:41:46 maybe you can enhance syntax to highlight tabs followed by spaces? Dec 12 18:42:52 JaMa: my pleasure Dec 12 18:43:39 JaMa: I have never done syntax highlighting work on vim, but what you say reminds me of a built-in feature of vim to highlight tabs (\t) Dec 12 18:44:14 or you are explicitly saying that the spaces followed by tabs be highlighted? Dec 12 18:45:22 sorry, tabs followed by spaces :) Dec 12 18:50:45 yes, that's always wrong according to OE style guildelines Dec 12 18:50:57 tabs themselves are OK when used in shell tasks Dec 12 18:52:09 JaMa: is there a tool that checks the validity of a bitbake file? Dec 12 18:52:15 bitbakelint, or bakelint etc Dec 12 18:53:12 instead of using vim to warn about the validity issues, I believe that a tool for checking the validity of a recipe accordingly to guidelines would be cool Dec 12 18:53:13 afaik no, but oe-classic has oe/contrib/oe-stylize.py Dec 12 18:53:54 true, there is patch to add some checks in create-pull-request script, I've proposed to check also for tabs-spaces there too Dec 12 18:54:30 but generic validity and styleguide script called from create-pull-request and manually if needed would be better Dec 12 18:54:56 when I was working on Pardus, we had a tool that checks the required items in package metadata and refuse the commit if not Dec 12 18:55:10 it was svn pre-commit hook on server side Dec 12 18:55:42 since oe uses git, making it sure in create-pull-request would be appropriate Dec 12 18:55:51 JaMa: what about single recipe updates? Dec 12 18:56:54 contributors need to be sure to run the validity tool before committing, I guess Dec 12 18:58:23 well contributors should test those recipes they are changing (build and runtime) so e.g. insane.bbclass should warn them about possible issues Dec 12 18:59:32 but I agree that some issues are easy enough to spot in commits itself so it's worth checking before sending pull request or patch Dec 12 19:07:33 I don't get why people use tab instead of 4 spaces Dec 12 19:07:49 I always use 4 spaces. Is there a specific reason for using tabs? Dec 12 19:25:03 I don't get it too.. you can find my emails about changing that (to unify shell tasks and python tasks) Dec 12 19:25:28 http://www.openembedded.org/wiki/Styleguide Dec 12 19:26:27 instead it was changed to say tabs for shell tasks (as yocto styleguilde has) http://www.openembedded.org/index.php?title=Styleguide&action=historysubmit&diff=4919&oldid=3443 Dec 12 19:49:39 can someone build devmem2? Dec 12 19:55:05 hrw: yes Dec 12 19:55:26 JaMa: it fails to put devmem2.c in workdir for me Dec 12 19:59:55 JaMa: what's your build machine? I feel like a turtle on my 1.2GHz cpu, 2gb ram laptop Dec 12 20:00:47 build machines... Dec 12 20:01:13 :) Dec 12 20:01:23 lucky Dec 12 20:01:41 8AMD FX(tm)-8120 here and few remotes Dec 12 20:02:15 intel core i7-2600k, 24GB ram Dec 12 20:04:33 * eren feels bad Dec 12 20:05:23 eren: my job for last 6 years was building software Dec 12 20:05:26 I'm desperately waiting for a server from my uni, which are being bought. God knows when they will be ready Dec 12 20:05:42 that's why fast build machine Dec 12 20:06:33 it always reminds me of: http://xkcd.com/303/ Dec 12 20:06:42 ;:)) Dec 12 20:06:47 common excuse Dec 12 20:07:29 yeah :) Dec 12 20:07:52 now I read the contiki guys, http://contiki-os.blogspot.com/2012/12/contiki-regression-tests-9-hardware.html Dec 12 20:09:18 anyway, I'm optimistic about the future. I will remember this days when I have blade servers interfaced with fiber-optic to storage unit :P Dec 12 20:10:13 * hrw off Dec 12 20:10:31 during night my machine will try to build all in meta-openembedded layers Dec 12 20:11:17 world? Dec 12 20:29:06 yes Dec 12 20:29:32 hrw: I'm building world with 19 layers, cca 54 failures after my last patchset Dec 12 20:41:51 I get the impression that OE is used in Sony Entertainment Dec 12 20:42:14 * eren referring to tim.bird@am.sony.com Dec 12 20:57:41 JaMa: I have more due to aarch64 Dec 12 20:58:07 eren: Tim Bird was also head of Consumer Electronics Linux Forum Dec 12 20:58:33 and first customer of my consulting company ;D Dec 12 20:58:57 hrw: oh, nice :) Dec 12 20:59:17 hrw: cool cloak of yours, btw Dec 12 20:59:30 linaro/hrw, hehe Dec 12 20:59:47 actually Tim usually complains that he's so far been unable to download and build w/ OE.. ;) Dec 12 20:59:50 eren: ? Dec 12 21:00:01 usually due to proxy or local machine configuration problems.. but he keeps commenting it's getting "better" Dec 12 21:00:02 ;) Dec 12 21:00:06 hrw: talking about your /whois output Dec 12 21:00:42 21:58 < eren> hrw: cool cloak of yours, btw Dec 12 21:00:54 eren: can you use other words? Dec 12 21:01:28 your whois output is: [hrw] (~hrw@linaro/hrw): Marcin Juszkiewicz Dec 12 21:01:57 linaro/hrw is called a cloak in freenode network. Mine is 'unaffiliated/eren' Dec 12 21:02:10 ah, ok now I got it Dec 12 21:02:41 a different cloak than unaffiliated makes you cool and elite Dec 12 21:02:42 :P Dec 12 21:03:48 eren: before Linaro I did not used any cloak Dec 12 21:04:13 there were options for named ones but I refused Dec 12 21:05:07 actually, I like seeing organizational cloak Dec 12 21:05:38 what I've seen so far in this channel are pdpc/supporter, rockbox/developer Dec 12 21:06:22 there were some problems in last months with cloak registration Dec 12 21:06:27 * hrw -> sleep Dec 12 21:06:32 have a nice rest of day Dec 12 21:06:42 good night Dec 12 21:32:20 This outlook express channel is awesome ;-) Dec 12 22:04:12 the poky danny branch fails to checkout completly on windows. Is that a intended "feature" or does nobody care? Dec 12 22:24:06 bitbake doesn't seem to love windows either :( Dec 12 22:24:41 who does.. Dec 12 22:25:49 mmm 99% of the pc users I am afraid Dec 12 22:26:38 I haven't seen single patch from them :) Dec 12 22:26:49 not sure if they love it thoug or are just married with it.... Dec 12 22:27:11 no so lets change that... Dec 12 22:27:19 lets hack in windows support Dec 12 22:27:37 it shouldn't be too difficult I guess... Dec 12 22:27:45 but perhaps I am wrong there Dec 12 22:28:46 since oe should be self contained,,, it is a matter of getting bitbake running (in an ideal world) Dec 12 22:30:54 well I think that getting bitbake to work on windows is way easier then most of recipes wich assume linux FS and tools available Dec 12 22:36:53 yes, you have a point there, it won't accidently get the host tools except for find and sort I guess Dec 12 22:39:06 bitbake should run out of the box on windows? Dec 12 22:39:50 isn't it pure python Dec 12 22:40:38 the first problems seems to be utils.py from bitbake. Using fcntl Dec 12 22:40:50 eren: yes I hoped so... Dec 12 22:41:32 not all parts of python are implemented on all platforms.. Dec 12 22:42:28 and yes, maybe I am doing something wrong.. (but I am running from msys / python 2.7) Dec 12 22:42:36 just install ubuntu. next, next, next, finish, reboot :) Dec 12 22:42:43 I had hoped that would work Dec 12 22:43:29 I haven't been used any windows for ~7 years Dec 12 22:43:31 eren: I have a working dual boot of windows 8 with ubuntu (even with this damn uefi) Dec 12 22:43:34 s/used/using Dec 12 22:43:44 wow, you have UEFI? Dec 12 22:43:50 this is the bonus excercise, get it working on windows as well Dec 12 22:44:08 UEFI should be pretty bad experience Dec 12 22:44:30 yes, it was, took ALLOT of time Dec 12 22:46:19 isn't it possible to just disable secure boot and get legacy booting procedure? Dec 12 22:46:45 I read that even reaching a bios menu was problematic and the key to press was not even referenced Dec 12 22:46:55 I see "Press F2 to enter" Dec 12 22:47:04 it was apperantly removed Dec 12 22:47:10 no, because windows 8 requires uefi and the new partition format Dec 12 22:47:37 I'm sorry but, WTF Dec 12 22:47:49 disable secure boot and you cannot boot win8? Dec 12 22:47:49 gparted refuses that and the windows 8 restore kicks in undoing what you did Dec 12 22:48:04 no not secure boot Dec 12 22:48:18 disable uefi and you cannot boot windows 8 Dec 12 22:48:46 UEFI is ok, no big issues Dec 12 22:48:53 unless secure boot is required Dec 12 22:49:32 UEFI is not ok for dual booting linux as it is on a Dell... Dec 12 22:49:43 here it works fine :P Dec 12 22:49:45 I thought that UEFI and secure boot are different Dec 12 22:49:52 win7 as well as win8 Dec 12 22:50:08 you can have uefi but do not use secure boot feature Dec 12 22:50:25 m$ products should be fine? Dec 12 22:50:38 * eren doesn't have any uefi hardware Dec 12 22:51:52 marketing secure boot feature just to lock users into one OS is really a nonsense. I don't recall any mainstream MBR viruses in 10 years Dec 12 22:52:21 it's harder then with linux (e.g. I don't know any way to switch to UEFI without reinstalling MS systems) Dec 12 22:52:46 JaMa: hm, I don't get it? Dec 12 22:52:49 with linux you can switch easily or keep both ways enabled Dec 12 22:53:28 eren: http://blog.flameeyes.eu/2012/12/debunking-efi-myths Dec 12 22:54:00 and yes "secure boot" is just one of uefi features and it can be disabled in some systems Dec 12 22:54:56 yes but there are far more issues Dec 12 22:55:21 _if_ you install e.g. ubuntu as non uefi Dec 12 22:55:44 the Dell rescue disk regards that as the rescue disk Dec 12 22:55:55 we're talking about an OS which refused to boot if it's not installed on first partition which is primary Dec 12 22:56:14 it was my experience with WinXP Dec 12 22:56:23 I don't know if it's the same with win7 Dec 12 22:57:45 so if for some reason (Linux installed) the Dell and or Windows 8 startup things are not sane, it starts linux.. and if that crashes (trying to use mdraid) it will start the windows reinstall Dec 12 22:58:18 anyway, not funny, will settle in time, works for the time being.... Dec 12 22:58:37 shim bootloader is signed with ms key, right? Dec 12 22:59:12 oh no, I'm confused I guess Dec 12 23:02:13 eren: secureboot is a non issue for people who know about pc, just disable it Dec 12 23:03:17 eren: the partition manager cannot install grub2, that is a bigger issue now, just takes time I guess,,, but is _really_ annoying Dec 12 23:04:35 and yes I have to get used to my new keyboard where ,,, and ... are messed up Dec 12 23:05:02 Dvorak? :) Dec 12 23:05:45 so back on topic, why does utlis.py need fcntl Dec 12 23:07:38 File "C:/msys/software/bpp3/system/poky/bitbake\lib\bb\utils.py", line 22, in Dec 12 23:08:04 import re, fcntl, os, string, stat, shutil, time Dec 12 23:08:21 ImportError: No module named fcntl Dec 12 23:08:39 is that the global lock, only one bitbake at a time? Dec 12 23:09:20 there are a number of file locks Dec 12 23:09:51 grep is your friend Dec 12 23:09:54 bah, so I really need to make this work on windows, instead of hacking it out? Dec 12 23:13:27 it's just not going to work on windows without a fair bit of work Dec 12 23:13:45 a linux VM would likely be a lot less painful route Dec 12 23:14:35 kergoth: yes thanks, I don't like vm's, I have ubuntu, xubuntu installed Dec 12 23:15:58 kergoth: I somehow hoped this would work on windows as well Dec 12 23:16:40 well actually I thinks it should Dec 12 23:16:57 good night all Dec 12 23:17:03 if you fix it, the patches would certainly be welcome Dec 12 23:17:11 to date no one has cared enough to invest the time Dec 12 23:17:31 yes, suprising to me... Dec 12 23:18:25 well, if I can I will have a look at it... Dec 12 23:18:45 but as said, also for me this is a bonus exercise Dec 12 23:19:26 but I will have to start with bitbake Dec 12 23:19:55 process.py seems to use the fnctl indirectly Dec 12 23:20:14 by the name I guess that is fundamental Dec 12 23:23:05 so lets park the windows till the weekend Dec 12 23:23:15 process.py is just a light wrapper around the python subprocess module, which may well go away at least in part once we require 2.7 Dec 12 23:23:29 a completely unrelated question Dec 12 23:25:10 on ubuntu the meta-toolchain-qte likes to build a cross compiler again as cross-canadian... what is so canadian about it it, and why can it not reuse the already build cross compiler? Dec 12 23:27:17 kergoth: is there a way to drop it already? Dec 12 23:28:02 kergoth: sorry I don't a single thing about python or bitbake, so this might be stupid Dec 12 23:28:57 s/ / know/ Dec 12 23:30:45 it's highly unlikely you're going to get bitbake to work on windows without knowing python :) Dec 12 23:31:19 kergoth: I will get to know it, it seems nice Dec 12 23:31:32 Guest87231: meta-toolchain* build a cross-compiler to run on SDKMACHINE. that is, it's capable of e.g. building an i686->arm sdk from an x86_64 host Dec 12 23:31:56 see also https://en.wikipedia.org/wiki/Cross_compiler#Canadian_Cross Dec 12 23:34:48 good, so we only need to add to build a windows8 host to be build on a mac for a linux target ;) Dec 12 23:36:01 but thanks, I set the target to 32 bit on 64 bit machine which likely triggers the candanian build... Dec 12 23:37:03 another one I am in the dark with is qt4-native, grep or bibtake -g doesn't tell me why it is build Dec 12 23:37:33 by what magic is qt4-native always included? Dec 12 23:38:51 At a guess, nativesdk-qt4-tools, given that's the only result from a grep for qt4-native Dec 12 23:38:58 (again, grep is your friend :) Dec 12 23:39:18 kergoth: yup grep that, remove it Dec 12 23:39:25 still builds Dec 12 23:39:45 it'll be in the .dot files, then Dec 12 23:40:37 ok, late here, will check it tomorrow.. Dec 12 23:41:27 even bumbing the PR doesn't help Dec 12 23:42:51 but perhaps there is a good reason to build it Dec 12 23:43:39 I am used to qt building native tools when cross-compiling Dec 12 23:43:39 it wouldn't be building if something didn't depend on it. the only remaining question is what depends on it, and is that behavior optional Dec 12 23:45:20 mmm, I couldn't find it in bitbake -g what pulled it in Dec 12 23:46:01 but qt needs native tools as well, so it might be needed, it is just the magic I don't like Dec 12 23:49:51 kergoth: thanks for your time Dec 12 23:49:57 got to go Dec 12 23:50:25 will look at this windows build this weekend **** ENDING LOGGING AT Thu Dec 13 02:59:58 2012