**** BEGIN LOGGING AT Fri Feb 24 03:00:00 2017 Feb 24 04:26:51 does meta-intel support i5 too ? Feb 24 04:27:16 in http://git.yoctoproject.org/cgit/cgit.cgi/meta-intel/tree/README it mentions i7 Feb 24 06:11:10 Does anyone knows where is the toolchain installed ? Feb 24 09:04:23 I Feb 24 09:04:30 d Feb 24 09:06:28 I wanted to try update to rt-kernel to I added the following in local.conf: #PREFERRED_PROVIDER_virtual/kernel ?= "linux-yocto" #PREFERRED_VERSION_linux-yocto ?= "4.4%" PREFERRED_PROVIDER_virtual/kernel ?= "linux-yocto-rt" PREFERRED_VERSION_linux-yocto-rt ?= "4.4%" Everything compiled well with bitbake core-image-sato-sdk Feb 24 09:06:43 I then decided to return to the regular linux by doing: PREFERRED_PROVIDER_virtual/kernel ?= "linux-yocto" PREFERRED_VERSION_linux-yocto ?= "4.4%" #PREFERRED_PROVIDER_virtual/kernel ?= "linux-yocto-rt" #PREFERRED_VERSION_linux-yocto-rt ?= "4.4%" But now I get errors as following: NOTE: Preparing RunQueue WARNING: /media/sde/yocto/poky/meta/recipes-sato/images/core-image-sato-sdk.bb.do_rootfs is tainted from a forced run NOTE: Executi Feb 24 09:07:16 File: '/media/sde/yocto/poky/meta/classes/license.bbclass', lineno: 48, function: license_create_manifest 0044: pkg_dic = {} 0045: for pkg in sorted(image_list_installed_packages(d)): 0046: pkg_info = os.path.join(d.getVar('PKGDATA_DIR', True), 0047: 'runtime-reverse', pkg) *** 0048: pkg_name = os.path.basename(os.readlink(pkg_info)) 0049: 0050: pkg Feb 24 09:29:15 any clue why a "mkfifo -m 0666 ${D}/usr/run/myfifo" in an do_install_append leady to a hanging do_package_qa in poky 2.0 which works in 2.1? Feb 24 10:11:33 rburton: I understand from the past that you dislike empty commas in PACKAGECONFIG[some] = ",,,". Can they all be omitted = "" if empty? Feb 24 10:17:55 Does anyone knows where is the kernel path ? Is it only in rootfs ? Feb 24 10:18:15 ranchu_: the kernel path? Feb 24 10:18:23 yes, Feb 24 10:18:32 what is a kernel path? Feb 24 10:18:36 yes Feb 24 10:19:30 I see in rootfs/usr/src. Is this the kernel source files ? Feb 24 10:19:57 sorry, please try to reformulate the question. i just cannot understand what you mean. Feb 24 10:20:55 LetoThe2nd - I have finished building image (no erros), I now try to find what is the path in /tmp for the linux kernel source code. Feb 24 10:24:12 Does anyone knows what is the path in /tmp for the linux kernel source code ? Feb 24 10:25:24 ranchu_: should be tmp/work-shared/$MACHINE/kernel-source Feb 24 10:27:21 Thanks ~ Feb 24 10:27:22 ! Feb 24 10:30:26 I don't have kernel-source . I have kernel-devsrc, which inside /1.0-r0$ contains: deploy-rpms license-destdir pkgdata sstate-install-packagedata sstate-install-package_qa sstate-install-package_write_rpm sstate-install-populate_lic temp Feb 24 10:31:31 Why I don't have kernel-source ? Feb 24 10:32:01 no idea. do you try to use external source? or are you using some ancient or morphed version of OE? Feb 24 10:33:54 no. it's latest. krogoth from meta-intel. Feb 24 10:34:37 krogoth and latest is already contradictory :-P and no idea about meta-intel, but i doubt they'd do heavy amgic there. Feb 24 10:34:58 simple build "bitbake core-image-sato-sdk" Feb 24 10:35:42 oops, I wander why it is so difficult.... Feb 24 10:35:51 What did I do wrong ?? Feb 24 10:35:56 well i just checked on my builds, being one krogoth/arm and one master/x86, and both behave exactly like i said Feb 24 10:37:51 do you also have tmp/work-shared/$MACHINE/kernel-drvsrc ? I ask becuase I really don't get what wrong...... Feb 24 10:38:10 nope, i don't have that. Feb 24 12:43:42 Heya! Feb 24 12:44:02 Did anyone ever had this error: | sudo: no tty present and no askpass program specified Feb 24 13:33:28 Hi everyone ! Has anyone had troubles with virtual/kernel -c menuconfig don't saving the conf ? I'm trying to add blutooth USB drivers but when I quit then reenter menuconfig all my config dissapeared. An idea ? Feb 24 13:36:33 One solution can be making a kernel receipe with a DEFCONFIG, but I don't know the name of the parameters I have to set to "y" Feb 24 13:44:00 hi all Feb 24 13:44:34 how can I make it so that a variable is not expanded during image construction phase but rather at execution? Feb 24 13:45:34 does the yocto build directory contain an unpacked rootfs somewhere that I can just export via NFS? Feb 24 13:45:40 I can fix it from poky's image.bbclass by doing a delVar on that particular variable, just like it's done there with DATETIME and TMPDIR Feb 24 13:46:01 but I would like to be able to do the same thing from my image recipe instead Feb 24 13:46:59 yes cdleonard under build/tmp/work/ Feb 24 13:47:16 look for a directory name containing your machine name at the beginning of its name Feb 24 13:47:31 inside that you should have your image recipe directory Feb 24 13:47:50 and browse that directory and you will find a rootfs directory Feb 24 13:48:51 there is something with $DISTRO in it's name, found it Feb 24 13:49:25 Anyone who knows how to fix: no tty present and no askpass program specified? Feb 24 14:09:42 problem is that I have an IMAGE_CMD_ which contains a certain value of a var, calculated based on a __anynomous function, but then the corresponding do_image_ function which has inlined the IMAGE_CMD_ does not contain the same variable value Feb 24 14:10:37 it contains the old value of that variable, the value it had before the calculation done in the __anonymous function from the IMAGE_CMD_ function Feb 24 14:11:33 how do I make sure that my recipe _anonymous function runs before the image.bbclass _anonymous function? Feb 24 14:47:25 no clue? Feb 24 15:26:35 Hi everyone ! Has anyone had troubles with virtual/kernel -c menuconfig don't saving the conf ? I'm trying to add blutooth USB drivers but when I quit then reenter menuconfig all my config dissapeared. An idea ? Feb 24 15:26:39 rburton: the strange kvm issue with dnf patchset seems to be fixed Feb 24 15:28:11 the image tests were testing package management by installing and deinstalling psplash, and somehow the system doesn't like if that's done rapidly Feb 24 15:28:25 I switched the tests to using safer packages as installation/removal targets Feb 24 15:29:01 cool Feb 24 15:30:52 rburton: there are more failures though, so I'll try to fix those too, and then we can do another AB pass Feb 24 15:31:02 this one was the most head-scratching though Feb 24 15:53:56 anyone knows the path to toolchain ? Feb 24 15:59:20 is it just me or is it not supported to export the rootfs from /tmp directly? I get some daemons failing at startup Feb 24 16:18:22 Does anyone knows the path to toolchain ? Feb 24 16:29:44 so there is no way currently to make a certain variable from an image to not get expanded when constructing the do_image_ function? Feb 24 16:31:02 oh jayis, i am going to need your laptop this weekend Feb 24 16:31:26 for like >30 min. i just want to see how long something will take to run Feb 24 16:44:09 First time irc user, is this the correct place to get help on yocto related problems? Feb 24 16:47:37 yes Feb 24 17:14:27 I see rburton active :) ; do you have any ideas related to fixing the issue I stated above? Feb 24 17:37:15 RP, do you happen to have a spare minute? Feb 24 17:37:40 gartin: you can ask a question and I can try and answer Feb 24 17:38:14 how can I make it so that a variable is not expanded during image construction phase but rather at execution? Feb 24 17:38:53 gartin: which execution do you mean? Feb 24 17:39:05 I've asked above; I can repaste if you can't see my previous questions Feb 24 17:39:30 thing is my do_image_ function differs from my IMAGE_CMD_ function Feb 24 17:40:15 gartin: I do have that in scollback Feb 24 17:40:22 in the sense that there's a variable I change in a recipe file using an __anonymous function and it shows up as expected in the IMAGE_CMD_ function, however in the do_image_ function it does not show up correctly Feb 24 17:41:11 in the do_image_ function it has the value before the processing done on it in the __anonymous function Feb 24 17:41:12 gartin: the problem is likely that those do_image_ functions are autogenerated Feb 24 17:41:18 yes Feb 24 17:41:31 is there a workaround for that? Feb 24 17:41:51 I can fix this by doing a delVar in image.bbclass Feb 24 17:41:54 but it's a no go Feb 24 17:42:11 gartin: can you arrange for your anonymous python to be included before the inherit of the image class? Feb 24 17:42:32 I *think* that would "fix" the ordering Feb 24 17:42:38 let me try that Feb 24 17:55:27 hello Feb 24 17:56:10 I need some help setting up a kernel recipe Feb 24 17:56:19 halstead: FYI centos7 on the new AB hung last night, just killed three build jobs which had been sitting idle without even starting bitbake for 21 hours Feb 24 17:56:45 rburton: halstead might be distracted atm Feb 24 17:57:08 is this the right place to get some help on yocto? Feb 24 17:58:27 rburton: I will have to look into it later. Do you suspect any cause in particular? Feb 24 17:59:22 doesn't look like I can; my __anonymous function is in a bbappend of a image recipe which itself includes core-image-minimal.bb RP Feb 24 17:59:33 halstead: no, haven't looked closely yet. just refired a buildset to see what happens. i'll let you get back to devday now :) Feb 24 18:04:24 RP, guess there's no way to "uninclude" a certain recipe from my bbappend so I can include it only later on? Feb 24 18:04:31 gartin: hmm, that would be a way to fix it... Feb 24 18:05:10 what way? Feb 24 18:05:17 gartin: no, there isn't a way to uninclude Feb 24 18:05:33 gartin: our replies overlapped, same there isn;t a way to change to order as I think it would fix it Feb 24 18:06:46 I could just remove the core-image-minimal.bb include from the main image .bb file and only include in my .bbappend after the __anonymous function; but thing is there are lots of others consumers of that .bb and it would look messy to have to do this for each of them Feb 24 18:07:39 gartin: the only other thing I can think of is to move your code to an event handler Feb 24 18:07:56 there should be an event that occurs before the anon python Feb 24 18:08:03 you could make your change there Feb 24 18:09:01 do you know off the top of your head where I could look for seeing how this is done currently in the poky code? any example existing? Feb 24 18:09:38 gartin: examples of events? Feb 24 18:10:48 The event I was thinking of is RecipePreFinalise Feb 24 18:10:52 yeah, on how to hook it up to the __anon python code Feb 24 18:11:11 RP, thanks I'll look up this event Feb 24 18:11:20 thanks for all the info Feb 24 18:11:52 gartin: take a look at multilib.bbclass as an example, your code can be simpler though Feb 24 18:12:24 RP, will do, thanks again Feb 24 18:26:47 RP, thank you so much, adding the code in an event handler like you suggested seems to have done it Feb 24 18:28:10 gartin: glad it helped Feb 24 18:46:33 Hello, I am trying to add opencv for python 2.7 within my yocto image and running into an issue where during the do_rootfs task it cannot find python-opencv Feb 24 18:47:11 I created a bbappends file to change the PACKAGECONFIG for opencv to python2 instead of python3 but I now run into this error Feb 24 18:56:32 someone sent a pseudo patch for posix_acl_default behavior, I haven't got a good way to test it right now, but I've tentatively merged a slightly-modified form into master. Feb 24 18:58:59 testing is over-rated. ;-) Feb 24 19:01:23 any tips on how to switch opencv to build for python2 instead of python3? Feb 24 19:04:32 uninstall python3 ? ;-) Feb 24 19:06:20 currently meta_oe/recipies-support/opencv/opencv_3.2.bb builds for python3 I added a bbappends file to change the packageconfig to python2 instead and now it is not able to fing the package during do_rootfs Feb 24 19:26:49 when are bbappend files processed in building a recipie, are they processed before or after the main bb file and how can I check to make sure that the changes to the PACKAGECONFIG I am making in my bbappends file are being made Feb 24 19:27:34 .bb is loaded first, then the .bbappends are loaded Feb 24 19:27:50 so a PACKAGECONFIG = ... line in a .bbappend will replace one specified in .bb Feb 24 19:28:43 would that also effect other lines such as ${@bb.utils.contains('PACKAGECONFIG', 'python3', 'python3', 'python', d)} or would they remain at the start determined by the pre bbappend PACKAGECONFIG Feb 24 19:30:45 bitbake -e yourrecipe to examine the final values of the variables Feb 24 19:39:27 well, im lost it appears the recipie is using python3 still in some places such as PYTHON_SITEPACKAGES_DIR even when I tell it to use python2 in the PACKAGECONFIG Feb 24 19:40:11 Hello friends. I am using the Freescale Community BSPs to build a Yocto image for an iMx6 board. I've gotten the core-image-minimal and core-image-x11 built and working using the dizzy branch. Now I am trying to get chromium into my build and I get the error 'no recipe available'. This is my first yocto project and I'm a little lost. Any questions Feb 24 19:40:11 or feedback appreciated. Feb 24 19:40:46 Majuk: search the layer index. Feb 24 19:40:56 https://layers.openembedded.org Feb 24 19:42:22 Majuk: you can look for recipes here https://layers.openembedded.org/layerindex/branch/master/recipes/ like kergroth said Feb 24 19:42:26 kergoth I cloned the meta-browser git. Tried both dora and jethro branches (as there is no dizzy branch) Feb 24 19:42:46 and you added it to BBLAYERS? Feb 24 19:42:54 Yessir. Feb 24 19:43:01 and the exact error is what? Feb 24 19:43:15 theoretically, I should be able to set the packageconfig for https://layers.openembedded.org/layerindex/recipe/39865/ to python2 instead of python3 correct? Feb 24 19:43:24 https://thepasteb.in/p/RghnErXPMPkHz Feb 24 19:44:08 https://thepasteb.in/p/GZhWkJ9yVmzfV Feb 24 19:44:14 Sorry, first one didn't get the whole thing Feb 24 19:44:38 That's using the dora branch of meta-browser. Feb 24 19:48:04 I'm guessing it's an issue that meta-fsl-arm has a different version of Chromium than meta-browser, but I don't know how to resolve that. Feb 24 19:48:16 or even how to confirm that's the issue. Feb 24 19:53:41 Ok, using the jethro branch of meta-browser, now it only complains about the gnome-settings-daemon recipe Feb 24 19:57:18 Specifically everything compiles but I get : Computing transaction...error: Can't install opencv-3.2+git0+70bbf17b13-r0@cortexa9hf_neon: no package provides python-opencv Feb 24 19:58:03 although with python2 in the PACKAGECONFIG it should be creating python-opencv Feb 24 20:01:59 Alright, one step forward, one step back. Added '${BSPDIR}/sources/meta-openembedded/meta-gnome' to my bblayers. Now it can find all the recipes, but I got a new error. Feb 24 20:02:21 whats the error Feb 24 20:02:27 https://thepasteb.in/p/Elh1LWjVRxPHm Feb 24 20:03:34 looks like you need the recipe libav Feb 24 20:03:55 https://layers.openembedded.org/layerindex/branch/master/recipes/?q=libav Feb 24 20:05:49 Nah, I have it, but I need to whitelist it Feb 24 20:05:59 ERROR: libav was skipped: because it has a restricted license not whitelisted in LICENSE_FLAGS_WHITELIST Feb 24 20:10:48 ...or maybe you're right. >.< Feb 24 20:14:45 im begining to think for some reason or anouther that the opencv recipie even though it has a config option for python2 cannot actually run with python2 Feb 24 20:15:01 Ive been trying to get it to work for a day or two now Feb 24 20:15:09 Nope, I was right, needed to whitelist "commercial" in local.conf Feb 24 20:15:41 It's building! I'm a wizard, Harry! Feb 24 20:15:52 Thanks for your time guys. Feb 24 20:16:57 nice Feb 24 20:30:55 i am running it right now to check again that it builds correctly with python3 but I really need python2 Feb 24 20:31:23 how do you switch from pyton 2 and 3? Feb 24 20:34:20 I created a bbappends file that contains Feb 24 20:34:45 PACKAGECONFIG = "python2 eigen jpeg png tiff v4l libv4l gstreamer samples tbb gphoto2 \ ${@bb.utils.contains("DISTRO_FEATURES", "x11", "gtk", "", d)} \ ${@bb.utils.contains("LICENSE_FLAGS_WHITELIST", "commercial", "libav", "", d)}" Feb 24 20:35:02 this is for opencv Feb 24 20:36:39 oh you meant for opencv Feb 24 20:36:50 i thought you meant for the distro Feb 24 20:38:05 it is specific to opencv both python2 and python3 succcessufly are installed in the distro Feb 24 20:41:25 to be more specific it is the recipie in meta-oe/recipes-support/opencv/opencv_3.2.bb Feb 24 20:50:18 is there a way to contact anyone who works on the support recipies in meta-oe to see if it is possible to switch opencv to python2? Feb 24 20:55:29 Does anyone knows where I can find toolchain in the tree ? Feb 24 20:57:32 ranchu: sysroots/[host architecture]/usr/bin/[target triple]/ Feb 24 20:58:11 bitbake -c devshell Feb 24 20:58:15 which gcc Feb 24 20:58:25 or that :) Feb 24 20:58:29 ;-) Feb 24 20:59:22 thanks, I'm checking now. Feb 24 21:03:11 rburton - I have this in this folder: user@user-VirtualBox:/media/sde/yocto/kro/poky/build/tmp/sysroots/intel-corei7-64/usr/bin$ ls crossscripts g-ir-annotation-tool g-ir-compiler g-ir-compiler-wrapper g-ir-generate g-ir-scanner g-ir-scanner-lddwrapper g-ir-scanner-qemuwrapper g-ir-scanner-wrapper Feb 24 21:03:13 what is target triple > Feb 24 21:03:14 ? Feb 24 21:03:34 ranchu: that's the target sysroot, not the host sysrot Feb 24 21:04:02 where you have intel-corei7-64, you want something like x86-64 Feb 24 21:04:14 where is the host sysroots ? Feb 24 21:04:22 ^ Feb 24 21:05:36 isn't "build/tmp/sysroots" is the host sysroot ? Feb 24 21:05:49 tmp/build/sysroots is the directory where all the sysroots are kept Feb 24 21:05:59 you have the native sysroot for stuff you can actually run Feb 24 21:06:05 and the target sysroots for stuff specific to each machine Feb 24 21:06:10 if you want gcc, its in the native sysroot Feb 24 21:06:29 a good question is "why do you care", as the right path will be in $PATH inside your recipe build Feb 24 21:09:05 ./x86_64-nativesdk-pokysdk-linux/opt/poky/2.1.2/sysroots/x86_64-pokysdk-linux/ Feb 24 21:09:08 Right ? Feb 24 21:09:54 oh, SDK Feb 24 21:09:57 you didnt say sdk Feb 24 21:10:16 seriously, what paulg is the best way if you're using OE/bitbake Feb 24 21:12:57 I got this with paulg idea: build/tmp/work/corei7-64-poky-lin ux/openssh/7.1p2-r0/openssh-7.1p2 Feb 24 21:15:29 so toolchain is in SDK , Right ? build/tmp/sysroots/x86_64-nativesdk-pokysdk-linux/opt/poky/ ? Feb 24 21:16:00 openssh? Feb 24 21:16:07 in a toolchain path? Feb 24 21:16:16 * paulg scratches head Feb 24 21:16:19 build/tmp/sysroots/x86_64-nativesdk-pokysdk-linux/opt/poky/ Feb 24 21:16:59 paulg - I did " bitbake -c devshell " as you suggested. Feb 24 21:17:11 devshell openssh Feb 24 21:17:58 so opencv will build correctly with the python3 PACKAGECONFIG but not with python2 any thoughts Feb 24 21:17:59 rburton - I care becuase I want to allow application staff to build application with toolchain Feb 24 21:18:41 What the best way to give application staff a development environment for our target (which is intel actually) ? Feb 24 21:19:06 I am familliar with toolchain , if there is something better please tell me. Feb 24 21:24:57 inkspell4: I am not familiar with opencv but it appears it is possibly to use python 2.7 Feb 24 21:25:01 http://www.pyimagesearch.com/2015/06/22/install-opencv-3-0-and-python-2-7-on-ubuntu/ Feb 24 21:25:34 "Update (3 January 2016): You can replace the 3.0.0 version with whatever the current release is..." Feb 24 21:25:54 have you seen this, does any apply to the error you are seeing? Feb 24 21:26:38 it does work with python2.7 but the recipie in meta-oe does not appear to accept the change from python3 to python2 in the PACKAGECONFIG Feb 24 21:27:10 I'm not sure I can't find in that path the gcc, and gdb binary. Any idea where's the toolchain (toolchain for host) ? Feb 24 21:27:18 Computing transaction...error: Can't install opencv-3.2+git0+70bbf17b13-r0@cortexa9hf_neon: no package provides python-opencv Feb 24 21:27:38 inkspell4: ah, sorry my bad, I didn't realize you were thinking the other way around Feb 24 21:28:12 inkspell4: that is at rootfs creation time you are seeing that msg? Feb 24 21:28:27 yes Feb 24 21:30:12 * marka follows, so if you look at the opencv /packages_split there is no python-opencv Feb 24 21:30:29 because, as you say it doesn't respect the PACKAGECONFIG Feb 24 21:30:37 * marka finally catches up to you Feb 24 21:31:31 is there a fix for this that I could easily implement? I am new to yocto (less than a week) and somewhat over my head on this one Feb 24 21:31:48 inkspell4: sorry, no sage advice then, do a devshell and run ../temp/run.do_configure Feb 24 21:32:29 you will have to walk through the configure process to sort out why it decides not to build the python hooks Feb 24 21:32:32 devshell? Feb 24 21:32:41 bitbake opencv -c devshell Feb 24 21:33:07 this will take you to a shell with the env setup for cross build... Feb 24 21:33:26 it is your best friend when debugging package configure/build/packaging issues Feb 24 21:33:39 bitbake opencv -e Feb 24 21:34:00 ok, is there also any way to make the maintainers of the recipe aware? Feb 24 21:34:02 is another friend, you can use this to ensure you are getting what you expect for variables like PACKAGECONFIG Feb 24 21:34:38 inkspell4: you can do a git history on the recipe and possibly contact them, but first best to understand a bit of what might be happenning Feb 24 21:34:57 ie. do the legwork to start Feb 24 21:35:30 folks here and such are much more apt to help when some of the issue is sorted out first Feb 24 21:36:11 best I can do as I have never looked at opencv Feb 24 21:39:02 ok, thank you for all your help Feb 24 21:39:19 I after invoking poky-glibc-x86_64-meta-toolchain-corei7-64-toolchain-2.1.2.sh I get in /opt/poky the toolchain, but not libraries, where is libpthread, librt ? Feb 24 21:41:26 inkspell4: np, sorry I was not more help Feb 24 22:11:35 So I compiled a new Yocto image with chrome on it. Now my problem is it appears any tab that tries to generate sound (videos or music [soudcloud]) causes the current tab to crash. Feb 24 22:12:18 I got a couple of warnings regarding libav and codecs-ffmpeg that I'm not understanding and I think they're related. Feb 24 22:12:57 https://thepasteb.in/p/Vmh03lny1lXt4 Feb 24 22:13:09 Any questions or feedback appreciated. Feb 24 22:17:47 according to old mailing list traffic, that should be only a performance issue, not a crash-inducing issue. are you able to get any kind of coredump? Feb 24 22:18:57 seebs: yea I was reading the same things you were. No core dump, the program itself keeps running, but I get that 'Aw, Snap!' page. Haven't found a way to get better info out of a tab/page crashing. Feb 24 22:20:03 No clue. There might be a setting somewhere for that, or it might matter if you use ulimit to enable core dumps; I know sometimes systems are configured by default to prevent actual core dumps. Feb 24 22:21:50 'ulimit -c 0' for some reason is common.. :P Feb 24 22:21:58 that will prevent core dumps Feb 24 22:22:12 Such joy, such pain, to get chromium-imx working and telling me that it's ready to give me hardware accelerated video decoding only to get derailed by an audio issue. Feb 24 22:23:01 My system doesn't have ulimit on it. Feb 24 22:24:52 there's probably a way to make the syscall directly, not sure. Feb 24 22:24:57 but congrats on getting the video working! Feb 24 22:25:26 lol, I mean, it SAYS it's working, but I can't prove it because the audio crashes the player! Feb 24 22:25:26 :D Feb 24 22:26:19 1 step forward, 1 step back. Every thread I'm finding regarding better logging of a tab 'Aw Snap' crash ends with "Oh, I found my problem some other way, google should really include some way to log these crashes" Feb 24 23:12:57 Does anyone know how to switch a build that utilizes python from finding the python3 to python2 interpreter? Feb 24 23:14:31 specifically with respect to a DEPENDS = "python' Feb 25 01:19:32 Hello, does anyone know of a good tutorial on creating a bbappend file to change the PACKAGECONFIG? **** ENDING LOGGING AT Sat Feb 25 03:00:01 2017