**** BEGIN LOGGING AT Thu Dec 19 02:59:59 2013 Dec 19 04:18:07 How can I remove something from MACHINE_FEATURES without resorting to editing the machine conf file? Is there an equivalent from EXTRA_MACHINE_FEATURES for removing? Dec 19 08:25:34 Can any body know the use of UBOOT_MACHINE ? Dec 19 08:31:43 ramose: I think it defines the U-BOOT configuration, since this doesn't necessarily need to change for different MACHINEs, UBOOT_MACHINE can stay the same while MACHINE changes for example Dec 19 08:34:39 Dennis__: Would it gives the name of config file?Is it U-boot specific config file? Dec 19 08:40:17 Try the following command from the yocto directory to find where the variable shows up $: grep UBOOT_MACHINE -r meta* Dec 19 08:55:06 hi all, Im wondering how can I choose gcc 4.6 instead of gcc 4.8 in dora ? Is there a variable to set in the local.conf ? Dec 19 09:01:56 ok, PREFERRED_VERSION_gcc = "4.x.x" Dec 19 09:02:23 but still don't know if it still working or deprecated Dec 19 09:53:53 morning all Dec 19 12:19:44 preferred version 3.2 of linux-libc-headers not available (for item linux-libc-headers-dev) Dec 19 12:19:51 How do I solve this? Dec 19 12:20:01 Google is not very helpful at this point.. Dec 19 12:23:59 kbart: people shouldn't be setting PREFERRED_VERSION_linux-libc-headers nor providing their own linux-libc-headers recipe (unless they really know what they are doing) Dec 19 12:24:16 kbart: if you're using a layer that does either of those things, go and hit the layer maintainer with a stick until they remove it Dec 19 12:25:00 arago_prefs.inc set it Dec 19 12:25:12 http://patchwork.openembedded.org/patch/57955/ Dec 19 12:25:27 is it a good idea to remove "PREFERRED_VERSION_linux-libc-headers_am335x-evm = "3.2"" ? Dec 19 12:25:47 I would suggest trying that yes Dec 19 12:25:57 denix: ^ Dec 19 12:27:05 ok, thank bluelightning, I'll try Dec 19 12:57:45 If I want to increase layer's priority (it must be the topmost) do I increase or decrease BBFILE_PRIORITY? Dec 19 13:06:08 BBFILE_PRIORITY only works if a file exists with the same name in 2 different layers, it then takes the file with higher priority Dec 19 13:11:49 It's about PREFERRED_VERSION_, lower layer (arago) defines one version of package, but I want to use another. I define PREFERRED_VERSION_ in my layer,, but it gets overwritten. How do I solve this? Dec 19 13:17:42 you can use a .bbappend file for the recipe that defines preferred_version_ and set your own in there, this overwrites the lower layer Dec 19 13:21:08 the problem is, that this PREFERRED_VERSION is defined in layer conf file. As far as I know, there no .bbappend files for them. Dec 19 13:21:41 and if I define PREFERRED_VERSION in my .bbappend or conf, it gets overwritten. Dec 19 13:23:41 kbart: PREFERRED_VERSION can't work in a bb/bbappend Dec 19 13:24:47 kbart: you can probably force it with _forcevariable but really arago should not be doing this if it's leading to problems Dec 19 13:27:27 I see. Thanks again. Dec 19 13:28:05 I guess the problem is that my kernel (3.2) is too old for this Arago release, probably the best would be to use older one that matches my kernel Dec 19 13:35:17 kbart: but hang on, linux-libc-headers does not need to match the kernel version - is that what it's trying to do? Dec 19 13:36:59 Not, it's another problem..;] Dec 19 13:37:20 I'm now trying to get libgles working Dec 19 13:38:08 I've got 4.09 working, but now it's defined to 5.01 and I can't overwrite it Dec 19 13:40:20 ultimately if you're fighting against the distro configuration it's a sign you should have your own Dec 19 13:47:50 Hm, I was thinking about that.. but it looks a messy job, setup all these toolchains etc.. is it worth it? I'm actually satisfied by Arago, I only need to use different kernel and change version of some packages Dec 19 14:14:41 Why does devshell now invoke as the root user? It used to run as the same user as the bitbake (e.g. tom) in Bernard-5.0.1 Dec 19 14:19:06 Is there a way to change a configuration of a recipe from another recipe ? Perhaps by using a global variable in core-image-test.bb and the package-recipe.bb is reading that ? Dec 19 14:19:35 hello everybody Dec 19 14:20:10 I managed to compile the default meta-ivi image, and now I am trying to build the meta-toolchain-sdk with it Dec 19 14:20:19 but I get the following error http://pastebin.com/ncXSeXD9 Dec 19 14:20:33 how to debug a bbappend recipe that patch a configuration file ? the patch is not applied Dec 19 14:20:35 I'm pretty new to yocto, any clue how to debug and/or fix this ? Dec 19 14:25:20 bluelightning: kergoth remember prior speaking of a qt4-x11-free issue where it does not build webkit? I tracked it down to getQMakeConf3's gawk statement in the gcc version detection Dec 19 14:49:15 any documentation out there to apply a patch to modify a configuration file ? ( I want to uncomment # mod_fastcgi in /etc/lighttpd.conf ) ? Dec 19 14:49:23 boucman_work: there are several similar fixes in the oe-core history, git log then search for "iles/directories were installed but not shipped" Dec 19 14:50:06 boucman_work: you have to either add the files explicitly or remove them before the QA check happens. Not sure which is right in your case. Dec 19 14:51:23 vmeson: thx, i'll git log a little and come back here after that... Dec 19 15:16:41 I'm having issues with libffi and it's include directory Dec 19 15:16:51 jamvm is trying to find ffi.h Dec 19 15:17:10 weebet: well, you don't need to patch it since that file is one supplied directly next to the recipe - just create a bbappend in your own custom layer to add your own version Dec 19 15:17:20 which is in the directory ./x86_64-linux/usr/lib/libffi-3.0.13/include/ffi.h Dec 19 15:17:57 rather than the more usual /usr/include, should I be hard coding an additional include path in the recipe? Dec 19 15:18:15 or is it possible that pkg-config might be able to pull out the correct include dir? Dec 19 15:20:16 thank you again bluelightning ! Dec 19 15:22:58 ah hah, a pkg-config hack in CFLAGS solved it Dec 19 15:34:26 Anyone have an idea why gettext-native sleeps forever in do_configure? Appears to be waiting for something? Dec 19 15:37:11 T0mW: anything interesting if you tail -f on log.do_configure for that recipe? Dec 19 15:39:18 I've seen this before... Dec 19 15:40:19 wasn't there a bug with the nvidia driver that was causing a SIGnal from the sleep to be corrupted? Dec 19 15:41:09 If I set the BBTHREADS + make -j flags to one thread, it progresses, but very slowly. If I set it to 8 threads, it locks forever. Dec 19 15:41:52 IIRC, I patched the automake config script, at one time, to skip certain tests involving sleeping. Dec 19 15:43:45 bluelightning: yes, it is the nvidia bug Dec 19 15:44:23 T0mW: that is a strange one... Dec 19 15:45:10 bluelightning: it really is, I had to google like mad to find some reference to it. this was a few weeks ago when I was updating another project to Dora. Dec 19 15:46:04 I'm running Nvidia GTX-650 video here with the proprietary driver, so, I've the bug. Dec 19 15:47:34 T0mW: out of interest do you know if the bug has been acknowledged/fixed by nvidia? Dec 19 15:50:03 bluelightning: naw, did not drill down that far. There was like one message on a site that someone identified the problem. I tried later drivers but did not resolve the issue. so, I patched gettext (bbappend) to skip those tests and force the proper result. Dec 19 16:02:08 bluelightning: check this link: http://forums-lb.gentoo.org/viewtopic-t-974600.html?sid=eb1c649ae513495183f1ad03c3a4ee19 Dec 19 16:02:56 bluelightning: that is where I got the info on the issue, then was able devise a work-around. Dec 19 16:14:38 DISTRO_FEATURES_append = " pam" leading white space give me build error while bitbake Dec 19 16:17:09 There are 8 patches of mt9v032(gumstix-camera) sensor in media-tree which are not included to my linux-gumstix. Is it a good idea to add those patches? or just work with what I have? Dec 19 16:21:07 bluelightning: you shouldn't tell what people should do and shouldn't :) Dec 19 16:21:25 denix: we can make pretty strong recommendations Dec 19 16:22:24 denix: what would you have told kbart instead? Dec 19 16:22:52 bluelightning: arago is a distro and not a bsp Dec 19 16:23:34 I'm not sure anyone suggested otherwise... Dec 19 16:23:42 bluelightning: plus we have a very wide range of kernel versions for different platforms - from 2.6.32 to 3.12 Dec 19 16:24:39 so, setting pref-version is rather required in our case Dec 19 16:25:24 providing a different linux-libc-headers per platform is still not advisable, for the reasons Richard outlined in the link I posted above Dec 19 16:25:40 making any suggestions or recommendations w/o knowing all the details is rather ill advised Dec 19 16:25:47 DISTRO_FEATURES_append = " pam" leading white space give me build error while bitbake.but DISTRO_FEATURES_append = "pam" works ok,why it is so? Dec 19 16:26:21 denix: ok, well next time I will direct enquiries to you then... ;) Dec 19 16:28:28 ramose: without a space is definitely wrong... what is the error you get with one? Dec 19 16:29:05 bluelightning: I understand Richard's point. but read Khem's comments in that discussion. he's quite rights and that's what I do in arago Dec 19 16:34:56 "However one case where it showed up was when building udev > 164 with kernels where accept4 call was not wired for arm e.g. since udev looked up definition of SOCK_CLOEXEC which it found but that 2.6.32 kernel really did not support it." Dec 19 16:35:40 bluelightning: getting error while patching shadow.automake-1.11.patch Dec 19 16:35:53 is it expected? Dec 19 16:37:50 bluelightning: and btw, the table can be rotated - why would kbart need to change pref-version for linux-libc-headers anyway? Dec 19 16:39:13 ramose: errors are never expected, what are the errors? Dec 19 16:45:41 patch is rejected while patching man/ru/Makefile.am file Dec 19 16:46:28 bluelightning: in the bbappend recipe to use another lighttpd.conf, what should be inside ? Only : princ := "${@int(PRINC) + 3}" \ SRC_URI += "file://lighttpd.conf" ? or I need something more ? Dec 19 16:59:34 how does the openembedded-core mailing list work? I just submit a patch and it gets applied with no discussion? Dec 19 16:59:42 is there any confirmation whether patch is gonna be applied or not? Dec 19 16:59:55 it's reviewed on the list Dec 19 17:00:08 if it's OK it's applied, if not you'll get comments in reply Dec 19 17:00:43 cool Dec 19 17:04:23 Xz: you generally have to watch the repo to see if it got merged, and occasionally patches get missed without discussion. it's not ideal but it (mostly) works. Dec 19 17:07:07 rburton: is there any schedule for merging? like every Friday or sth similar? Dec 19 17:07:42 Xz: stuff generally merges twice a week, after review+build tests in master_under_test on the autobuilder Dec 19 17:09:43 Xz: there is a mailing list for commit notifications as well: openembedded-commits@lists.openembedded.org Dec 19 17:10:27 ndec: ok, so monitoring that one actually gives an overview of what exactly gets merged, is that correct? Dec 19 17:11:05 it gives you anything which is committed. the email is sent once the commit is pushed. Dec 19 17:37:53 Can't install oracle-jse-jdk-arm-1.8.0-u120r0@cortexa8hf_vfp_neon: no package provides libX11.so.6 Dec 19 17:38:12 any hints how to solve? locally working fine, autobuilder throws error :( Dec 19 17:38:26 ccube: nothing depends on libx11 so it isn't getting built? Dec 19 17:39:39 mh Dec 19 17:39:46 yes Dec 19 17:39:50 no gui stuff here Dec 19 17:39:57 but why then error? Dec 19 17:41:07 because the jdk wants libx11 Dec 19 17:41:07 rburton, should my java package depend on it? Dont see any reason for. Dec 19 17:41:19 mh Dec 19 17:41:21 the error says that the jdk package has a dependency on libx11 Dec 19 17:41:28 java crap Dec 19 17:41:58 yes, but if it depends on, then it should build it Dec 19 17:42:01 you should be able to configure your way out of that one... Dec 19 17:42:37 ccube: as it's the oracle jdk, was that built yourself or just a downloaded binary? Dec 19 17:42:43 the jdk doesn't *need* x/gtk unless you want to build java/swing stuff Dec 19 17:43:04 yes Dec 19 17:43:09 i dont want swing Dec 19 17:43:12 so no x11 Dec 19 17:43:18 is there a java virtual? Dec 19 17:43:36 fwiw, debian's openjdk packages has a no-x11 version. Dec 19 17:43:38 RPROVIDES_${PN} += "java2-runtime" Dec 19 17:43:40 ? Dec 19 17:43:41 i bet icedtea would satisfy the no-X dependency Dec 19 17:46:23 * mr_science needs a yocto build env here in the office... Dec 19 17:46:35 :D Dec 19 17:47:01 openjdk is building fine Dec 19 17:47:08 (with dirty workarounds) Dec 19 17:47:56 rburton: There are 8 patches of mt9v032(gumstix-camera) sensor in media-tree which are not included to my linux-gumstix. Is it a good idea to add those patches? or just work with what I have? Dec 19 17:47:58 should be able to pass it via extra oeconf Dec 19 17:47:58 what kind of vnc packages do i need for sanity tests? Dec 19 17:48:51 Zaif: igenuinely have no idea whatsoever Dec 19 17:49:22 rburton: haha ok. no problem Dec 19 17:49:39 ccube: what you actually need is an X server. xfake, tigervnc, a real one, doesn't matter. Dec 19 17:51:02 rburton, kk thx Dec 19 17:52:45 ccube: tigervnc is useful because when it totally breaks, you can vnc in and see Dec 19 17:57:08 kk, give it a try Dec 19 18:03:34 denix: fair enough; perhaps I was a little too hasty in responding Dec 19 18:10:19 Can't install oracle-jse-jdk-arm-1.8.0-u120r0@cortexa8hf_vfp_neon: no package provides libasound.so.2(ALSA_0.9.0rc4) :/ Dec 19 18:10:23 java sucks Dec 19 18:10:29 seems to be random errors Dec 19 18:13:25 weebet: you also need FILESEXTRAPATHS_prepend := "${THISDIR}/files:" Dec 19 18:13:52 weebet: this allows the build system to locate additional files in a "files" subdirectory next to your bbappend Dec 19 18:25:49 ccube: a long list of missing dependencies more like Dec 19 18:28:50 yeah, would be nice... Dec 19 18:29:11 I really wonder why the build just failing on one box Dec 19 18:29:24 ah Dec 19 18:29:27 now it works Dec 19 18:29:33 but sanity fails :D Dec 19 18:32:32 oops, didnt emerge the server :/ Dec 19 18:32:40 next try... :D Dec 19 18:53:04 does we need to inherit pkgconfig in every recipes ? Dec 19 19:01:58 weebet: only if the software being built by the recipe uses the "pkg-config" tool to find how to link to some of its dependencies (usually its configure script) Dec 19 19:05:00 thank you bluelightning Dec 19 19:16:34 aha. got it. Dec 19 19:17:10 qt4, 4.8, 0007-qmake_cxx_eval.patch Dec 19 19:23:15 WarheadsSE: hmm, IIRC that was a patch I wrote Dec 19 19:23:33 I do know it broke the ability to run configure outside of our environment, which is a bit bad... Dec 19 19:24:35 grep = bad. Dec 19 19:24:57 >> MAKE_CONF_COMPILER=`getQMakeConf "$XQMAKESPEC" | grep "^QMAKE_CXX[^_A-Z0-9]" | sed "s,.* *= *\(.*\)$,\1," | Dec 19 19:25:08 that should read Dec 19 19:25:33 +QMAKE_CONF_COMPILER=`getXQMakeConf QMAKE_CXX | sed "s,.* *= *\(.*\)$,\1," | sed -n 's/$[(]\([0-9a-zA-Z_]*\)[)]/$\1/pg' | tail -1` Dec 19 19:25:48 I think that has been changed recently, let me check Dec 19 19:26:12 well this is also sadly against a 4.8.4 that someone _here_ copied off 4.7?*? Dec 19 19:26:25 oh, well that might explain the difference Dec 19 19:26:31 but I looked at the dylan (which this should be based off) qt4 Dec 19 19:26:56 let me check what we have Dec 19 19:27:00 looking at this : https://github.com/openembedded/oe-core/blob/denzil/meta/recipes-qt/qt4/qt-4.8.0/qmake_cxx_eval.patch Dec 19 19:27:26 that seems like it was a layered on patch Dec 19 19:27:50 http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-qt/qt4/qt4-4.8.4/0016-configure-eval-QMAKE_CXX.patch?h=dylan Dec 19 19:27:54 that is the dylan version Dec 19 19:28:00 no, that's denzil .. I have been punching people for being behind. Dec 19 19:28:11 Yup Dec 19 19:28:16 * WarheadsSE eyeroll Dec 19 19:28:21 * WarheadsSE goes to kick developer. Dec 19 19:30:00 The following post-install scripts could not be run offline and rootfs is read-only: 100-openjdk-7-vm-zero Dec 19 19:30:19 any hints? Dec 19 19:30:40 I wanna keep the ro-fs Dec 19 19:30:45 (if possible) Dec 19 19:33:47 ccube: what's that script doing? Dec 19 19:35:20 dont really know Dec 19 19:35:24 cannot find it in my FS Dec 19 19:35:39 seems like openjdk pulls it in somewhere Dec 19 19:37:08 bluelightning: Yeah.. we have a release version that is still denzil based Dec 19 19:37:10 on my local box it is working, on autobuilder not :( Dec 19 19:37:22 WarheadsSE: understood Dec 19 19:37:25 So I'll probably just pull dylan into our repo for 4.8.4 Dec 19 19:37:34 er, dylans recipe Dec 19 19:37:48 vs our broken 4.8.4 made from 4.8.0 or some nonsense Dec 19 19:38:44 ccube: you should be able to see it under tmp///1.0-r0/rootfs/etc/rpm-postinsts/ Dec 19 19:39:39 WarheadsSE: good that you managed to track down the issue Dec 19 19:40:07 yeah, that was a days worth of pain Dec 19 19:43:22 WarheadsSE: heh, yeah, simple shi^H^H^H stuff always seems to suck up a day to find the solution. Hard stuff is easy. Dec 19 19:44:30 I was chasing why the USB keyboard would not work, turns out udev needed a volume mounted before it ran (time == 6 hours to resolve) Dec 19 19:46:08 obviously there will always be problems like this, but if anyone can suggest places where the build system doesn't catch and report errors/misconfiguration in an appropriate way, we'd really love to hear about it Dec 19 19:47:41 bluelightning, i dont see this folder structure somewhere Dec 19 19:47:53 it seems to be completely different on my builds Dec 19 19:49:31 ccube: what do you see currently? Dec 19 19:50:40 about thousand open terminal windows^^ Dec 19 19:52:27 ccube: right... well in the one where you have the error you pasted there should be the path to the work directory for the image, can you see that? Dec 19 19:52:53 yeah Dec 19 19:53:08 and you don't see rootfs/etc/rpm-postinsts/ under that? Dec 19 19:53:58 aah Dec 19 19:54:01 sure Dec 19 19:54:51 bluelightning, http://pastebin.com/Au03EQZs Dec 19 19:55:50 ccube: so, for some reason that is failing Dec 19 19:56:16 ccube: I can't tell why that would be, but the script's not that long, so I'm guessing it shouldn't be too hard to figure out Dec 19 19:56:36 T0mW: yeah.. we all have days like that Dec 19 19:56:45 ccube: if it helps, $D would be set to the full path to the rootfs directory when this script was called Dec 19 19:57:17 what I dont understand is why these build are fine on one box and failing on another. Dec 19 19:57:48 I'm not sure either, sorry Dec 19 19:59:04 bluelightning, thx :) Dec 19 20:17:48 I'm scratching my head over this one "Files/directories were installed but not shipped" Dec 19 20:18:25 files or directories were installed into the ${D} in do_install, but were not listed in the FILES_* directives used to package the files.. Dec 19 20:18:57 ok, I did have the dir in the FILES_${PN} though Dec 19 20:19:38 use bitbake -e and check that it's there and hasn't been overridden in some way Dec 19 21:31:08 Well, the file shows up in the image directory but it still complains about '... were installed but not shipped' Dec 19 21:31:26 This worked with Bernard-5.0.1, but not with Dora Dec 19 21:32:20 update one or more package defs? Dec 19 21:32:58 either that path is missing or the PN_files is too restrictive Dec 19 21:33:09 or maybe the order of packages is incorrect Dec 19 21:33:22 *order of things getting packaged Dec 19 21:35:53 Well, this is the linux-yocto-custom which gets a strange name, I'm assuming that 'FILES_${PN}' is correct? work directory struct looks odd: '...linux-yocto-custom/3.4.74+gitAUTOINC+c26a9a9684-r6/linux' Dec 19 21:36:11 not sure of what ${PN} would be Dec 19 21:36:50 Or maybe I should be using 'FILES_${PN}${PV}' ? Dec 19 21:39:19 hmm, maybe 'install -d ${PV}' and see what directory it creates within images Dec 19 21:42:26 T0mW: the value of PN is derived from the recipe name Dec 19 21:43:56 yeah, should be Dec 19 21:44:08 FILES_${PN} would be the correct variable to set/add to to influence what goes into the main package for the recipe Dec 19 21:47:55 ::w Dec 19 21:50:29 bluelightning: if you would please check this? http://pastebin.com/CRrSZ1PV Dec 19 21:51:08 bluelightning: this is a simple recipe technique that worked under Bernard. I just keep googling and come up with nothing. Dec 19 21:51:32 The do_install_append worked before. Dec 19 21:54:10 kernel builds and installs fine, but, I have a custom header within the kernel that I want to push out into the rootfs image inside /usr/include/linux/ Dec 19 21:56:56 which package do you want it in? ${PN}-dev ? Dec 19 21:57:38 in the general package, don't need it for dev Dec 19 21:58:20 I am building pkg-dev so if it can be triggered from that, ok. Dec 19 21:59:33 typically headers would go into the -dev package Dec 19 21:59:48 0: linux-yocto-custom-3.4.74+gitAUTOINC+c26a9a9684-r7 do_package (pid 7775) Dec 19 21:59:59 no dev extension? Dec 19 22:00:32 that's the recipe name rather than a package name (although there is almost always a main package with the same name as the recipe) Dec 19 22:01:41 nothing is working, I'm just throwing stuff into the recipe at this point. Dec 19 22:02:17 trying this now: FILES_${PN}-dev = "${includedir}/linux/sys3000.h" Dec 19 22:03:09 It has got to be so obvious that I just cannot see it. Dec 19 22:07:12 tomorrow is another day(tm) Dec 19 22:14:01 T0mW: hmm, FILES_${PN}-dev should already have /usr/include in its value, so it's odd that it wouldn't be picked up Dec 19 22:17:47 bluelightning: doesn't the kernel alter PACKAGES? Dec 19 22:23:43 RP: well, I just checked the output of bitbake -e linux-yocto to get that answer... Dec 19 22:24:45 oh, wait, you're right, no linux-yocto-dev Dec 19 22:25:45 T0mW: ${PN}-dev probably doesn't exist for the kernel, try FILES_kernel-dev += "${includedir}/linux" instead Dec 19 22:45:18 can someone tell me if yocto supports the at91sam9m10g45ek Dec 19 22:46:44 ermmm AT91SAM9G45 Dec 19 22:55:47 OutBackDingo: what exactly is that chip? Dec 19 22:55:53 armv7? Dec 19 22:59:07 mr_science: ARM926EJ-S [41069265] revision 5 (ARMv5TEJ), cr=00053177 Dec 19 22:59:49 Machine: Atmel AT91SAM9G45-EKES Dec 19 23:00:04 armv5t should be supported, so you would probably need to tune it yourself Dec 19 23:00:58 i did that for the rpi Dec 19 23:01:13 should be very similar i would think Dec 19 23:01:23 it says it was supportd by the kalios project which was now integrated into yocto but i see no references to MACHINE ?= "at91sam9g45ek" Dec 19 23:01:48 it wouldn't be a machine name Dec 19 23:02:16 my machine is still "raspberrypi" but i modified the tune stuff Dec 19 23:03:06 hang on a sec while i clone... Dec 19 23:03:49 look in conf/machine/include/ for some tune files Dec 19 23:04:09 eg, tune-arm1176jzfs.inc for rpi Dec 19 23:04:29 you might need to make one for AT91SAM9G45 Dec 19 23:05:20 * mr_science goes back to testing Dec 20 00:26:47 ok seriously confused...... i see kaeilos was integrated to yocto and i see oe-core and openembedded Dec 20 00:28:21 but i see no references to kaeilos in yocto, and kaeilos has board supprt according to http://www.at91.com/forum/viewtopic.php/f,12/t,18967/ Dec 20 00:33:20 i also see at91sam9g45ek.conf in https://github.com/openembedded/openembedded/tree/master/conf/machine Dec 20 00:33:30 * OutBackDingo is lost Dec 20 00:34:42 if i put MACHINE ?= "at91sam9g45ek" it says Please set a valid MACHINE in your local.conf or environment Dec 20 00:45:23 OutBackDingo: you would need to find or add the machine definition for that machine Dec 20 00:45:45 that would mean finding or creating a layer adding support for the machine Dec 20 00:48:43 bluelightning: yes i understand this, i followed http://www.kaeilos.com/download which stated KaeilOS is now integrated into Yocto Project as custom layer, so looking around yocto i see 0 references to it Dec 20 00:49:25 OutBackDingo: to say it's "integrated *into* the yocto project" isn't quite right Dec 20 00:49:30 but if i follow their "legacy" guide it actually accepts the board i have as MACHINE ?= "at91sam9g45ek" and starts the build, unfortunately Dec 20 00:49:39 it fails in bb Dec 20 00:49:40 I think what is meant is it's built on top of it Dec 20 00:50:11 i would have thought it to be a "meta" Dec 20 00:50:38 if you mean an additional layer, yes Dec 20 00:52:52 though i cant find a single reference to it Dec 20 01:01:07 you should probably ask mckoan about kaeilos, it's his project Dec 20 01:01:16 he'll be asleep now though Dec 20 01:15:16 bluelightning: thanks, at least its a lead to pursue :) Dec 20 01:16:18 OutBackDingo: FWIW, you may find support for a specific machine is available independent of a distro such as kaeilos Dec 20 01:16:38 at least, that's typically how things are now structured Dec 20 01:16:48 * bluelightning needs sleep **** ENDING LOGGING AT Fri Dec 20 03:00:00 2013