**** BEGIN LOGGING AT Tue Dec 05 03:00:01 2017 Dec 05 06:29:39 Hi, how can we create toolchain which can be used in windows ? Dec 05 06:29:46 or sdk for windows Dec 05 08:08:52 ranran, I would say meta-mingw : http://bfy.tw/FOXX, but I never tried Dec 05 08:09:20 ranran (but I already searched) Dec 05 08:09:57 aurele: he's gone, i would already have answered though. Dec 05 08:10:08 aurele: and i tried for morty lately, and it broke. Dec 05 08:12:08 LetoThe2nd, I feel so bad to be late... but if anyone come with some question he could find the answer (lets hope someone will search such a thing) Dec 05 08:14:18 aurele: technically you're totally right (no need to feel bad :-P) - unfortunately this channel is not logged publicly, and nowadays people only seem to search stackoverflow. Dec 05 08:15:22 Hi guys, I'm trying to compile Open Clovis (SAFplus) with yocto (rocko branch). But the meta layer (https://github.com/joaohf/meta-openclovis) is not updated. Can I get some help here? Thx. Dec 05 08:16:38 kosta_z: well, take over maintenance and update it :-P if you run into specific problems you can always ask and we'll try to help, but i don't think anybody here will step up and maintain the layer here. Dec 05 08:17:32 LetoThe2nd: Ok. I'll reproduce the specific problem and ask here. Thanks man! )) Dec 05 08:19:31 kosta_z: generally, try to put logs that are as informative as possible into a pastebin, and share the link here together with a short description. then its easiest for us to have a look. Dec 05 08:25:16 LetoThe2nd: Cool man. I got it. I'll be back here soon with the details. Dec 05 10:10:42 Hey, I've got an error in GCC 7.2 do_patch for ./poky/meta/recipes-devtools/gcc/gcc-7.2/0051-ARM-PR-82445-suppress-32-bit-aligned-ldrd-strd-peeph.patch (1 out of 2 hunks FAILED -- rejects in file gcc/testsuite/gcc.target/arm/peep-ldrd-1.c) Is it a known issue ? Dec 05 10:43:21 Hey, I've got an error in GCC 7.2 do_patch for ./poky/meta/recipes-devtools/gcc/gcc-7.2/0051-ARM-PR-82445-suppress-32-bit-aligned-ldrd-strd-peeph.patch (1 out of 2 hunks FAILED -- rejects in file gcc/testsuite/gcc.target/arm/peep-ldrd-1.c) Is it a known issue ? Dec 05 10:43:29 oops sorry flood Dec 05 10:48:17 nayfe: what oe-core, and do you have any BSPs that would be also patching gcc? Dec 05 10:50:12 awww MUTplosion all over the autobuilder :( Dec 05 10:50:15 rburton: using freescale bsp, but no gcc patch, seems to be that commit http://git.yoctoproject.org/cgit.cgi/poky/commit/?h=rocko&id=d164445477d57c18e93fd280ecccfa64b8394246 Dec 05 10:50:42 nayfe: reverted it to verify Dec 05 10:50:43 ? Dec 05 10:51:21 ok i'm trying Dec 05 10:56:20 rburton> compiles fine without this commit Dec 05 11:13:10 nayfe: please email the list with the consolidated pull where this landed, obviously a pretty bad problem. Dec 05 11:13:17 armpit: ^ Dec 05 11:13:27 RP: reports that rocko's gcc is bust Dec 05 11:13:33 (replicating now) Dec 05 11:16:20 I did some builds with updated rocko already and haven't see this one yet Dec 05 11:18:58 the AB should have noticed it obviously Dec 05 11:19:06 got a rocko running now Dec 05 11:39:55 this one also seems to be new with latest rocko update ERROR: QA Issue: ELF binary 'work/qemux86-oe-linux/go-dep/0.3.0-r0/packages-split/go-dep/usr/bin/dep' has relocations in .text Dec 05 11:43:31 nayfe: rocko builds fine here. double check the freescale bsp doesn't patch gcc? Dec 05 11:56:39 It seems VAR:= in a .bbclass is not overrideable by VAR= in local.conf. I suppose I need to do _VAR:="somethingimmediate" and then VAR?="${_VAR}" in the bbclass to allow it to be overrideable by local.conf, right? Dec 05 11:59:45 yeah VAR_DEFAULT in the class would give you the single expansion you're after but still overridable Dec 05 12:01:43 rburton: VAR_DEFAULT was a better name. Is this by convention? Dec 05 12:01:54 no just better than mystery underscores Dec 05 12:02:10 I completely agree, thanks Dec 05 12:08:43 I've inherited a Yocto project. When I run the image under qemu I do not have any network enabled other than loopback. I can see that this *can* be controlled by the presence of CONFIG_E1000E=y in a config file. Searching through the *idiotic number* of configuration files I can see several places where that value is set. BUT, I can't figure out a way to know that any of those configuration files are actually processed. Dec 05 12:09:02 How do I do that Dec 05 12:12:05 New news from stackoverflow: Which yocto is best for me Dec 05 12:25:05 Smit-Tay: go to kernel work directory, find the .config, read it Dec 05 12:52:37 rburton: all the not yet merged upgrade patches are here https://git.yoctoproject.org/cgit/cgit.cgi/poky-contrib/log/?h=akanavin/package-version-updates Dec 05 12:52:48 thanks kanavin1 Dec 05 12:52:49 rburton: I'll also send to oe-core in a moment Dec 05 12:59:09 rburton: adding musl support to AUH would be particularly useful, as it's a notorious source of not-even-tested-by-submitter problems Dec 05 13:07:04 if its doing builds a musl one would be invaluable Dec 05 13:10:53 rburton: it is, I backed off from the idea of removing build support altogether Dec 05 13:11:15 rburton: so I guess it can take the list of machines, and build for musl only for the first one in the list Dec 05 13:11:29 rburton: is TCLIBC=musl bitbake sufficient? Dec 05 13:19:02 good question Dec 05 13:19:17 i *think* TCLIBC gets pulled in from the environment Dec 05 13:19:40 rburton: I'll try in a sec, just need to finish the current build Dec 05 14:00:58 Hello there, I want to know a few things about how to configure yocto such that (1) a simple application (let's say a hello world C program from a new recipe) is run at boot time. (2) run a kernel module from a simple recipe at boot time and automatically build its "/dev/hello" interface file (which normally is done by mknod). Any example/tutorials regarding those questions? Thank you very much. Dec 05 14:04:11 lau_cris: 1) write an init script or systemd service (depending on what you're using for init) 2) the kernel module should be causing the node to appear magically thanks to udev. Dec 05 14:22:54 hi, what does RDEPENDS_${PN}_class-target do? I dont understand the *class-target* part. Dec 05 14:25:29 fl0v0: http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#ref-classes-nativesdk Dec 05 14:25:40 fl0v0: Sets the RDEPENDS_${PN} variable when building for the target architecture. nativesdk builds will be unaffected (as will native builds, but those don't go through packaging anyway) Dec 05 14:26:33 thank you! Dec 05 14:27:28 ah man, i should have search in the manual. just used google Dec 05 14:28:00 don't forget about the bitbake manual too. if it's not in the yocto manual it is usually there Dec 05 14:28:03 usually :) Dec 05 14:34:15 stephano: why are you up? Dec 05 14:37:12 rburton: i have a final exam this week, and a project due as well. Dec 05 14:37:19 coffee is my friend :) Dec 05 14:37:24 hooray coffee Dec 05 14:37:31 \o/ Dec 05 14:43:23 kanavin1: see ross/mutb for a better fix for psmisc Dec 05 14:46:08 rburton: I replicated that fix precisely? Dec 05 14:46:17 huh Dec 05 14:46:28 rburton: is it this? http://git.yoctoproject.org/cgit/cgit.cgi/poky-contrib/commit/?h=ross/mutb&id=e899dabe3a51d060c764bd6a804b174ab78ab9fc Dec 05 14:46:32 oh turns out i fetched in the wrong clone Dec 05 14:46:43 sorry :) Dec 05 14:47:15 rburton: git branch -D ross/mut* :) Dec 05 15:08:24 rburton: for gcc patch problem, to me there is an issue in patch http://git.yoctoproject.org/cgit.cgi/poky/plain/meta/recipes-devtools/gcc/gcc-7.2/0051-ARM-PR-82445-suppress-32-bit-aligned-ldrd-strd-peeph.patch Dec 05 15:09:03 hi all, I'm getting an error ...data_smart.py ; when I'm trying to bitbake some image... I never saw that before anybody have idea what is the issue or I did something wrong Dec 05 15:09:30 xtron: way more context needed Dec 05 15:10:26 nayfe: works for me so i guess the question is why does it break for you Dec 05 15:10:48 rburton: bitbake report "unable to parse ../../data_smart.py" I think this is some general issue.. ok i share the pastbin of this error Dec 05 15:10:55 rburton: in peep-ldrd-1.c patch, we replace "{ dg-final { scan-assembler "strd" } }" by "{ dg-final { scan-assembler "strd\\t" } }" then we copy peep-ldrd-1.c into peep-ldrd-2.c and replace "{ dg-final { scan-assembler "strd" } }" by "{ dg-final { scan-assembler-not "ldrd\\t" } }" except that it is already "strd\\t" by previous chunk Dec 05 15:11:14 xtron: sounds like you've got a corrupted data_smart.py, its part of bitbake Dec 05 15:12:18 rburton: maybe, but this file is not something we modify normally Dec 05 15:13:02 here is the pastebin https://pastebin.com/r0mCeBvH Dec 05 15:14:13 getVar() missing 1 required positional argument: 'expand' Dec 05 15:14:27 rburton: what is this? Dec 05 15:19:37 xtron: your bitbake needs you to always pass the expand variable, but your metadata does't Dec 05 15:21:23 maybe running new poky clone with an older bitbake? Dec 05 15:55:33 i'm using the yocto plugin in eclipse with an installed sdk from a yocto build, there are unresolved includes in most files, includes that were resolved when using a CDT managed project, how do I fix this in a yocto cmake project? Dec 05 15:57:21 YPTM: dial into: 1-800-262-0778 enter the attendees number: 88748961 Dec 05 15:57:32 YPTM: Stephen Joined Dec 05 15:57:41 YPTM: Saul is on Dec 05 15:58:47 YPTM: Stephano's here Dec 05 15:59:32 hello bitbake world Dec 05 16:00:20 YPTM: Richard joined Dec 05 16:00:45 YPTM: armin joined Dec 05 16:01:09 YPTM: Rob joined Dec 05 16:01:27 YPTM: bavery joined Dec 05 16:01:32 Chin Huat here Dec 05 16:01:35 YPTM: Patrick joined Dec 05 16:03:22 Aaron Chan joined Dec 05 16:03:59 YPTM: Randy joined. Dec 05 16:04:01 Did the YPTM access code change? Dec 05 16:04:03 YPTM: Mark is on Dec 05 16:04:47 JPEWhacker: dial into: 1-800-262-0778 enter the attendees number: 88748961 Dec 05 16:05:08 Ok, finally worked Dec 05 16:05:56 ah, yptm! Dec 05 16:06:12 Is it okay to join the YPTM call just to listen what is going on? Dec 05 16:06:21 yes.. meeting is open Dec 05 16:06:49 If you join, annouce yourself on the IRC. :) Dec 05 16:06:51 milindur: absolutely all are welcome: dial into: 1-800-262-0778 enter the attendees number: 88748961 Dec 05 16:07:00 YPTM ross joined, sorry i'm late Dec 05 16:07:09 YPTM: Joshua Watt here Dec 05 16:07:44 Richard is discussing the issues we are seeing in the build/test environment. Dec 05 16:08:11 kanavin1: can you rebase your devtool patches on top of paul's patchbomb this morning? you've less patches so he wins... Dec 05 16:08:34 YPTM: Christian joined Dec 05 16:10:05 YPTM wmills is here Dec 05 16:11:37 What are the common failures caused by memory fragments in KVM in AB test? Dec 05 16:11:58 aaronchu: qemu can't start with kvm allocation failure Dec 05 16:12:23 rburton: are they in a branch on poky-contrib? he posted them via oe-core-contrib, which I don't work with :-/ Dec 05 16:12:33 kanavin1: one sec can push Dec 05 16:13:10 kanavin1: ross/devtool Dec 05 16:15:07 aaronchu: https://autobuilder.yocto.io/builders/nightly-x86-64-lsb/builds/630/steps/Running%20Sanity%20Tests/logs/stdio Dec 05 16:15:15 aaronchu: specifically | qemu-system-x86_64: failed to initialize KVM: Cannot allocate memory Dec 05 16:15:55 its trying to do a 256kb kalloc which the kernel can't do as the free memory is fragmented too much Dec 05 16:18:10 This happens running on any distro? Dec 05 16:19:57 aaronchu: fedora 26 and suse tumbleweed only Dec 05 16:20:02 which suggests its a new kernel thing Dec 05 16:20:12 rburton: pushed to same branch akanavin/package-version-updates Dec 05 16:20:13 happens more on f26 in my experiene Dec 05 16:20:20 kanavin1: great thanks Dec 05 16:20:21 How do you reproduce the smart error? Dec 05 16:20:30 JPEWhacker: build images with rpm Dec 05 16:21:26 k, I', building morty on FC 27, I'll try a build quick to see if I can reproduce it Dec 05 16:22:14 nayfe: do you have PATCHTOOL set in your local.conf? (re gcc patch problem) Dec 05 16:23:36 RP, Caches dropped. Adding job to do it nightly but only when load is low. Dec 05 16:24:00 halstead: will it wait if the load is high? Dec 05 16:24:11 rburton: nop (bitbake -e => quilt) Dec 05 16:24:28 ive tried changing my FILESEXTRAPATHS_prepend to := "${THISDIR}/files", and also tried removing the line entirely, neither changes the result. Dec 05 16:24:32 halstead: i tend to fire builds before going to bed so what timezone is nightly in? :) Dec 05 16:24:51 rburton, That's my plan. Sleep until the load is low and run. Dec 05 16:24:59 cool Dec 05 16:24:59 halstead: not bothered about the load, I'd just do it Dec 05 16:25:43 rburton: i submitted a patch on oe-core http://lists.openembedded.org/pipermail/openembedded-core/2017-December/145283.html Dec 05 16:25:51 YPTM: is over Dec 05 16:25:57 nayfe: yes, just trying to understand why Dec 05 16:26:12 we see this now and again where everyone has a patch that applies apart from a few people Dec 05 16:26:18 must be some host difference Dec 05 16:26:23 nayfe: this builds for us but not for you and we'd really like to understand why. Which host distro are you using? Dec 05 16:26:27 rburton: is quilt doing stuff in parallel ? Dec 05 16:26:30 nayfe: nope Dec 05 16:26:45 we even build our own quilt Dec 05 16:26:46 nayfe: do you have any custom quilt/patch stuff in $HOME ? Dec 05 16:27:01 now that YPTM is over, what happened to libgcc?? Core image minimal (rocko) does not have one while sato image does Dec 05 16:27:24 wmills_: depends on what links to it? Dec 05 16:27:33 Are the autobuilders actual machines, or are they containers/vm? Dec 05 16:27:34 isnit it often statically linked? Dec 05 16:27:44 Does anyone knows how to share a sock proxy thru KVM from host? Dec 05 16:27:49 RP, alright. I was thinking I'd add to /etc/cron.daily. So rburton that means it would run around 17:45 UTC. Dec 05 16:27:51 I see some eabi_ functions in libc itself. Does it staticly pull in everything it needs from libgcc?? Dec 05 16:28:07 wmills_: it will only get pulled in if something links to it. Not all binaries need libgcc Dec 05 16:28:09 AB can run in containers/VM or host machines. Dec 05 16:28:58 halstead: that sounds fine Dec 05 16:29:00 Yes but before most things needed it yes? Was there a change in the way glibc is built? Dec 05 16:30:33 Probably a question for Khem, just curious anyway Dec 05 16:32:11 wmills_: http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=709aac9d25170b6a05d7904727bd7a06bb30f959 - back in 2012 we (and glibc) relaxed the requirements Dec 05 16:32:31 nayfe: can you pastebin the gcc-source-7.2.0 recipe's log.do_patch? Dec 05 16:32:47 (before your fix) Dec 05 16:34:13 RP rburton: my mistake, i had a old host patch version Dec 05 16:34:23 nayfe: how old? Dec 05 16:35:31 RP rburton: quite old, as I had issue in binary patchs with buildroot :p Dec 05 16:35:37 heh Dec 05 16:35:46 tell us a version and we can blacklist it at startup Dec 05 16:35:54 that is why I'm asking :) Dec 05 16:36:49 patch 2.6.1 :) Dec 05 16:36:55 MY GOD Dec 05 16:37:07 you do know thats eight years old right? :) Dec 05 16:37:26 nayfe: did you verify that upgrading to 2.7 fixed it? Dec 05 16:37:58 * rburton tries downgrading to see what happens Dec 05 16:38:20 rburton: I assume you're going to add a minimum version check? Dec 05 16:38:33 rburton: nop, i just removed local bin to current 2.7.5 Dec 05 16:38:58 and with it (current version) it patches without issue Dec 05 16:41:30 cool Dec 05 16:41:32 RP: yes Dec 05 16:43:32 rburton: thanks :) Dec 05 16:43:43 well of course it worked with 2.6.1 Dec 05 16:44:06 oh it didn't actually do anything, lets try again Dec 05 16:45:05 My FC27 build of morty with PACKAGE_CLASS = "package_rpm" was successful. Is there an example of the autobuilder failure somewhere? Dec 05 16:45:33 JPEWhacker: smart hangs, there's no logs of the error becase we have to go and kill the python Dec 05 16:45:46 its smart -> python -> libdb -> pthread Dec 05 16:46:27 RP: ok whats the trick to forcing gcc-source to refetch Dec 05 16:46:37 gcc-source-7.2.0 -C unpack isn't working :/ Dec 05 16:46:45 oh dinner, back on this later Dec 05 16:46:48 rburton: k, I guess I was looking more for the build setup so I can get closer to the build actually causing the problems than my own custom image Dec 05 16:47:42 ok, i have some info on my recipe, non-working example and then command to make it work https://pastebin.com/BNuCavSP Dec 05 16:48:35 rburton: I'd just "bitbake gcc-source-7.2.0 -c clean; bitbake gcc-source-7.2.0 -c patch" Dec 05 16:50:26 JPEWhacker: https://autobuilder.yocto.io/builders/nightly-packagemanagers/builds/262 Dec 05 16:50:53 JPEWhacker: https://autobuilder.yocto.io/builders/nightly-packagemanagers/builds/262/steps/CreateAutoConf/logs/stdio is the config it was using, you can then see the step that was killed after 28 hours Dec 05 16:51:13 JPEWhacker: I'd suspect it was in do_rootfs of core-image-sato-sdk Dec 05 17:18:12 any ideas how a patch can be successfully applied and yet not modify the files it is supposed to, but work properly and modify files when forced to run a second time? Dec 05 17:23:45 otavio: Hi, on my imx6dl custom board, I use adv739x TV encoder, and driver (video/fbdev/mxc/mxcfb_adv739x.c) uses fbdev stack, i suppose i need to rewrite it in mainline kernel to use new DRM/IPU stack ? Dec 05 17:26:42 right so patch<2.7 doesn't really support git-style diffs at all Dec 05 17:29:29 rburton: indeed, but somehow i get 2.6.1 to work until yesterday :) Dec 05 17:29:51 nayfe: that new patch uses some git-style features Dec 05 17:30:04 we've had this a few times in the past so i'm happy with just telling people not to use a patch from 2010 Dec 05 17:30:12 :D Dec 05 17:30:36 seems fair Dec 05 17:39:37 hmm, i'm unsure on how to write a recipe for a python package that uses pip and virtualenv Dec 05 17:40:39 basically, i've got a script that depends on Flask, and a requirements.txt Dec 05 17:43:52 yes of course patch changed how it outputs version between 2.6 and 2.7 Dec 05 17:47:01 rburton: of course it did Dec 05 17:54:47 nayfe: patches on the list to add a new sanity check. would be appreciated if you could hide the new patch you installed, apply the patches, and delete conf/sanity_info Dec 05 17:54:57 nayfe: then running bitbake should tell you your patch is too old Dec 05 18:02:06 is bitbake 1.30.0 really old? i thought the krogoth branch of yocto had a new bitbake Dec 05 18:21:21 hmm, if i want pip in my configure task, what do i have to do? python3-pip-native does not seem to exdist Dec 05 18:21:24 *exist Dec 05 18:42:50 uglyoldbob: older releases have their matching bitbakes Dec 05 18:50:29 Is morty building on Fedora 26? Dec 05 18:55:12 rburton> i'll do it tomorrow keep in touch Dec 05 19:19:22 join #qt Dec 05 19:19:26 ...oops. Dec 05 19:21:56 hmm.... http should be a default module in python3.... right? = Dec 05 19:22:27 because i installed python3-werkzeug, in yocto and when importing it i get No module named 'http' Dec 05 19:23:36 ok i created a bug at https://bugzilla.yoctoproject.org/show_bug.cgi?id=12397 Dec 05 19:23:37 Bug 12397: normal, Undecided, ---, richard.purdie, NEW , patch_do_patch fails to detect patch failure Dec 05 19:24:06 peacememories: our python modules are packaged in a highly granular fashion, many modules/packages get their own binary packages. either install python3-modules, or better yet see exactly which packages you need for its deps and add them to the rdepends for werkzeug Dec 05 19:25:45 kergoth then it's still a bit weird that the werkzeug recipe doesn't itself depend on any of it, right> Dec 05 19:25:59 many, many, many recipes are missing rdeps on the needed python modules Dec 05 19:26:12 it's an issue that needs fixing, on a per-recipe basis Dec 05 19:26:17 also, i can't find python3-modules anywhere Dec 05 19:26:34 the python-modules/python3-modules package depends on all the modules in the standard library Dec 05 19:27:51 somewhere around here i have a text file where i ran devtool create on every python recipe i could find, comparing the resulting rdeps against those in the existing recipe, to see which ones needed fixing Dec 05 19:30:35 hmm, where did i put that Dec 05 19:31:46 https://gist.github.com/kergoth/7d2e9f6c7d0fd759d7f2#file-check-python-deps-txt .. course i haven't run it in ages, not current, and required manual review for each Dec 05 19:41:20 peacememories: python3-modules is an empty package created by the python3 recipe that simply has deps on all the module packages Dec 05 19:43:35 New news from stackoverflow: How to pass compile time parameter for Yocto project? Dec 05 19:43:48 and the http modules is also generated by python3, i assume? Dec 05 19:48:17 peacememories: yes Dec 05 19:49:48 can't find it thought >.< python3-modules i can find via oe-pkgdata-utils, but not anything else^^' Dec 05 19:50:01 peacememories: it's in the python3-netutils package Dec 05 19:51:04 peacememories: FYI I just found that by doing oe-pkgdata-util find-path *python*http* Dec 05 19:58:26 i tried oe-pkgdata-util show-pkgs | grep python3-http, i guess that wasn't the right thing to do^^ Dec 05 20:01:00 hmm, i'm not finding the same thing Dec 05 20:02:51 python3-netutils doesn't exist anywhere for me Dec 05 20:03:17 i'm on morty, maybe that's a difference? Dec 05 20:04:44 also, python3-modules does not seem to provide http in my case O.o Dec 05 20:13:57 hmm Dec 05 20:28:26 huh, i got it to work by moving the rdepends from my recipe to a bbappend for flask Dec 05 20:28:26 not sure why this changes anything in this case but... sure? Dec 05 20:28:28 thanks for your help :) Dec 05 20:33:09 only thing is that it takes an eternity to start^^ Dec 05 21:11:59 how do i access files added to SRC_URI in do_install? Dec 05 21:12:07 ah nvm, now it worked^^ Dec 05 21:12:32 they're unpacked into WORKDIR Dec 05 21:29:51 I'm seeing some odd compiler errors with poky/master that I didn't see in morty. I'm using gcc 6.x in both cases. This simple test case builds in morty, but not git master: https://pastebin.com/bvMm7Euy Dec 05 21:31:35 srwalter: what error do you see? Dec 05 21:33:19 bluelightning: https://pastebin.com/KyDq5Urv Dec 05 22:50:43 I'm seeing some odd compiler errors with poky/master that I didn't see in morty. I'm using gcc 6.x in both cases. This simple test case builds in pyro, but not git master: https://pastebin.com/bvMm7Euy Dec 05 22:50:57 khem: rewitt suggested you might be able to help Dec 06 00:57:39 waht's the difference between IMAGE_INSTALL and IMAGE_INSTALL_append ? Dec 06 01:27:18 learningc: same variable, but IMAGE_INSTALL_append adds the value to the end Dec 06 01:28:00 note that _append does not add a space in between the existing and appended value so you need to include that in the appended value Dec 06 01:28:56 I see, Can I only use IMAGE_INSTALL_append (no IMAGE_INSTALL in local.conf at all) ? Dec 06 01:45:24 learningc: well, if you set it outright there you will probably find that it'll either be overwritten by what the image recipe sets Dec 06 01:45:50 learningc: really if you want custom image contents you are best off creating your own custom image recipe, then you have full control Dec 06 01:46:56 s/either// **** ENDING LOGGING AT Wed Dec 06 02:59:59 2017