**** BEGIN LOGGING AT Fri Nov 04 03:00:00 2016 Nov 04 03:06:00 to do it temporarily is trivial, can use the menuconfig task for that. bitbake -c menuconfig virtual/kernel. how you make it persistent depends on the kernel recipe you're using, which dpeends on the bsp you're using. Nov 04 08:53:32 oO ... hmm basically a pretty general question ... where to stop with build dependencies .. if you have a simple "helloworld" cmake project would you need to specify cmake, make and gcc as dependency? and when which one? have seen cmake.bbclass adds cmake-native why not nativesdk-cmake to get it in the SDK as well? Nov 04 09:02:06 mdnneo: gcc/binutils/and so on are part of the default depends Nov 04 09:02:22 mdnneo: as you say, inherit the class and you'll get cmake-native for you Nov 04 09:02:38 what goes into a SDK is unrelated and depending on nativesdk-cmake would be wrong Nov 04 09:03:31 so again to get make and cmake in the SDK I would need to explicit specify either in the image or in some custom sdk packagegroup Nov 04 09:03:58 if you want nativesdk-cmake in your sdk then you need to add it explicitly, yes. Nov 04 09:04:55 (by adding nativesdk-cmake to TOOLCHAIN_HOST_TASK in your image recipe) Nov 04 09:08:43 so there is no way of just put something in if it is needed ... lets say the helloworld has some make target to generate doxygen docs I would need to figure out per image if the package is in and if nativesdk-doxygen is needed? Nov 04 09:09:15 why would you need doxygen because helloworld is in the image? Nov 04 09:09:28 the SDK doesn't contain the tooling to rebuild every package that goes into the image Nov 04 09:10:04 its your SDK so you decide what goes into it. by default you get a compiler and some very common tools (see packagegroup-sdk-host). Nov 04 09:10:41 ok ... so the SDK is basically in the end kind of independent of the image ... at least the host part Nov 04 09:10:51 the host part, yes Nov 04 09:10:59 well Nov 04 09:11:12 the image has control over what is in the host part, but it's not generated from the content of the image itself Nov 04 09:12:12 ok so lets say you can control whats in in the image recipe but its independent of the contained packages Nov 04 09:45:11 hi everybody, I've got a problem while trying to build an existing recipe on a new machine. Nov 04 09:45:50 I'm using daisy and when I'm trying to build the same image (all git revs are the same) on a new debian x86_64 machine i get the following error: Nov 04 09:46:18 ERROR: OE-core's config sanity checker detected a potential misconfiguration. Nov 04 09:46:24 Either fix the cause of this error or at your own risk disable the checker (see sanity.conf). Nov 04 09:46:30 Following is the list of potential problems / advisories: Nov 04 09:46:39 TUNE_ARCH is unset. Please ensure your MACHINE configuration includes a valid tune configuration file which will set this correctly. Nov 04 09:46:47 Error, the PACKAGE_ARCHS variable does not contain TUNE_PKGARCH (None).Toolchain tunings invalid: Nov 04 09:46:53 Tuning 'cortexa9thf-vfpv3d16' has no defined features, and cannot be used. Nov 04 09:47:02 ERROR: Execution of event handler 'check_sanity_eventhandler' failed Nov 04 09:47:07 ERROR: Command execution failed: Exited with 1 Nov 04 09:47:11 if your layers are all identical, it sounds like your local.conf isn't Nov 04 09:48:58 rburton: I've also thought that, but I've checked the local conf "a million" times now and also copied it directly... still the same error Nov 04 09:49:34 double-checked the layer/sha listing on startup is identical? Nov 04 09:51:22 The only difference in local.conf is the "DL_DIR", but that (IMHO) shouldn't affect it... Nov 04 09:52:30 do you mean the listing on startup of bitbake? This doesn't occur because the sanity checker errors are printed and bitbake exists before that... Nov 04 09:54:50 I compared all layers listed in bblayers.conf with "git log -1 --oneline" and the matched. Nov 04 09:55:27 Also "git status" said "Your branch is up-to-date with 'origin/BRANCH'" on both machines Nov 04 09:55:41 have you also triple-quadruple checked bblayers.conf, not only local.conf? :-) Nov 04 09:56:11 And the origin remote of the git repos match too ;-) Nov 04 09:59:03 LetoThe2nd: yes.. only the paths in BBLAYERS & BBLAYERS_NON_REMOVABLE are different, but point to the "same" local git repository... Nov 04 10:01:01 g0hl1n: no layers missing, or paths looking right but actually being invalid? Nov 04 10:06:10 LetoThe2nd: just checked it again. All of the four lines in BBLAYERS are present and pointing to the "same" local git repository and at least "ls"ing and "cd"ing them works Nov 04 10:08:34 g0hl1n: well there ought to be a difference, somewhere. or bad karma. Nov 04 10:09:23 g0hl1n: I guess it's most likely the poky repo. Those tune setting are in poky/meta/conf/machine I think Nov 04 10:09:41 another idea would be local modifications that do not show up in the git log Nov 04 10:09:58 g0hl1n: by the way, what do you mean by "same" git repo? Nov 04 10:14:51 To me it looks like vfpv3d16 was added after daisy, see http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/conf/machine/include/arm/arch-armv7a.inc?h=daisy vs http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/conf/machine/include/arm/arch-armv7a.inc?h=master Nov 04 10:15:03 g0hl1n: ^^ Nov 04 10:16:17 erbo: I've got this machine definition in my meta layer. otherwise it wouldn't work on my current/old machine ;-) Nov 04 10:20:50 Hi everyone... I tried to build an image for my Beaglebone black yesterday using Toaster. I got it to build but now I can not find the .hddimg file. It seems to be missing. Or am I missing something here? Nov 04 10:23:39 LetoThe2nd: I've also checked for local modification using "git status". There are none. And just for your understanding: Until now I've done the builds on my dev machine. Currently I'm trying to build on a separate server (to speed up the build). On that server a script is checking out the defined tags/branches of all needed layers before building. Therefore I'm pretty sure there are no local modifications to those repos. Nov 04 10:24:34 LetoThe2nd: With "same" repo I mean it's the same remote with the same revision in a separate local checkout ;-) Nov 04 10:25:26 cart_man: your images go to /build/tmp/deploy/images/ Nov 04 10:26:14 cart_man: Err.. Ignore the leading slash .. it doesn't go to your / (root).. copy-paste error Nov 04 10:26:44 g0hl1n: i still guess the script is missing some step that you maybe subconsiously would do when setting up everything manually Nov 04 10:27:18 JoiF: I am in that file...theres a bunch of stuff there but no hddimg file Nov 04 10:27:50 cart_man: I don't even know what an "hddimg" file is Nov 04 10:28:27 cart_man: But you kernel, root-filesystem and everything else is there Nov 04 10:28:31 JoiF: its the file that you can copy to your SSD Nov 04 10:28:54 JoiF:Yea but in sepearte arts... I dont know how to build it from all those files Nov 04 10:29:16 cart_man: IMAGE_FSTYPES contains said hddimg? Nov 04 10:29:56 LetoThe2nd: huh? Nov 04 10:30:18 cart_man: what huh? Nov 04 10:30:26 is IMAGE_FSTYPES a flag or something ? Nov 04 10:30:51 cart_man: nope, a configuration variable that tells bitbkae which image artifacts you desire to reciev in the end. Nov 04 10:31:06 including typos, of course! Nov 04 10:31:14 LeteThe2nd: Ohhh ... hmm ill quickly check Nov 04 10:31:42 cart_man: http://www.jumpnowtek.com/beaglebone/BeagleBone-Systems-with-Yocto.html Nov 04 10:31:47 cart_man: bitbake -e your-image-thing | grep IMAGE_FSTYPES Nov 04 10:32:46 cart_man: Instructions are for krogoth though .. morty should be pretty much the same thing Nov 04 10:33:33 hddimg is not really a file system Nov 04 10:34:02 jku: sure, but the variable is called like that. Nov 04 10:34:38 yeah but I don't think hddimg is going to be in that list Nov 04 10:35:58 depends on the machine configuration, it certainly could be pulled it. so in a nutshell: "it depends" Nov 04 10:46:24 Ok so I used core-image-sato-sdk and it built fine twice before.,.. SO I added hddimg flag and now it stops with 2 errors -> Nothing PROVIDES 'syslinux'syslinux was skipped: incompatible with host arm-poky-linux-gnueabi (not in COMPATIBLE_HOST) Nov 04 10:46:43 Required build target 'core-image-sato-sdk' has no buildable providers. Nov 04 10:46:43 Missing or unbuildable dependency chain was: ['core-image-sato-sdk', 'syslinux'] Nov 04 10:48:00 cart_man: obviously the hddimg image type is defined to be using syslinux. Nov 04 10:48:13 cart_man: what brings you the idea of using that for the bbb anyways? Nov 04 10:48:40 cart_man: IIRC meta-ti brings everything along for creating the sd card, but as "live" and flasher images. Nov 04 10:48:42 LethoThe2nd: Well largely because I dont know how to create the image using anything else... that ons was easy. You just DD it onto a disc and there you go Nov 04 10:52:53 cart_man: let me guess, knowig that procedure from some other platform? Nov 04 10:54:27 cart_man: my suggestion would be to have a good look at https://www.yoctoproject.org/downloads/bsps/morty22/beaglebone Nov 04 10:54:43 cart_man: it describes the procedure for a bbb sd card in good detail Nov 04 10:55:52 LetoThe2nd: thanks, I will check that (after lunch) ;-) Nov 04 11:42:20 hello, I am having a bad time making NFS boot work with my Yocto build using systemd Nov 04 11:43:20 I get failed unit dependencies during boot and eventually end up in the emergency mode prompt Nov 04 11:44:18 Booting via network used to work fine using sysvinit for the Yocto distro Nov 04 11:50:07 http://pastebin.com/w7DSMWTP Nov 04 11:50:20 How to compile recipe with license in INCOMPATIBLE_LICENSE list defined by other layer Nov 04 11:57:43 Siva_: what sets INCOMPATIBLE_LICENSE? It's usually something you would only get from your own configuration, not from 3rd party layers... Nov 04 12:00:13 eduardas_m, log in and look why the services failed? In the first error message you can read: See 'systemctl status systemd-modules-load.service' for details. Nov 04 12:00:37 @jku: Thanks for reply. I'm porting some distro layer on hikey, so that layer sets the conf. Nov 04 12:04:15 Siva_: I guess you need to remove that setting from the layer or avoid building recipes with incompatible license Nov 04 12:07:57 Cubi_, this unit is probably not related to my problem, it fails even uring normal boot from sd card and all else still functions normally in that case Nov 04 12:08:41 I get only 2 failed units during normal boot from sd card, though Nov 04 12:09:03 during booting from network I get many more Nov 04 12:11:01 @jku problem is I need to remove it from my layer Nov 04 12:11:53 @jku I can't modify that distro, that why looking for options like WHITELIST Nov 04 12:15:02 hi guys Nov 04 12:15:42 I built a SDK via populate_sdk to work on a software that is installed on my image Nov 04 12:16:15 If I want to insert meta-qt5 into my Yocto build do I add it into the local.conf file ? Nov 04 12:16:17 but this software needs to run a command called ngrestcg, which is native Nov 04 12:16:40 but it is not going on x86 sysroot sdk, it exists only on core2 sysroot Nov 04 12:17:07 cart_man: the layer needs to be present in bblayers conf, and you need to define what you want from it Nov 04 12:17:17 How can I built a sdk with this binary on x86 sysroot? Nov 04 12:18:08 CTtpollard : Ok but do I have to download it from GIT first ? Or do just add /home/machine/Yocto/poky/meta-qt5 ? Nov 04 12:18:35 you need to provide a path to the source layer yes, where ever that is located is optional Nov 04 12:19:43 CTtpollard : Ok soo I can download from git and insert it into a folder called /home/machine/Yocto/poky/meta-qt5 AND then add that exact filepath to my bblayer.conf ? Nov 04 12:20:18 indeed Nov 04 12:20:33 CTtpollard: Ok and then its included... nothing else to do ? Nov 04 12:21:51 cart_man: well nothing in your build will change, unless you instruct something to be built Nov 04 12:25:04 Cttpollard : Not quite sure what that means ? .. You mean I have to let it know to build it in conf file or so? Nov 04 12:26:08 That just makes bitbake aware of the layer, you the need to instruct it to build something that makes use of the layer Nov 04 12:26:50 Cttpollard: Well I want it to build qt5 ... but I dont know where to actually insert XX += qt5 Nov 04 12:26:58 which file actually needs to be told Nov 04 12:29:05 #cart_man to include a layer in build do bitbake-layers add-layer /meta-qt5 Nov 04 12:35:55 Siva : my bitbake-layers show-recipes seem to crash with an error ParseError Nov 04 12:37:03 Its hacing this error because of the meta-qt5 layer in inserted into the folder Nov 04 12:38:41 you've probably got a misconfigured bblayers.conf Nov 04 12:39:11 CTtpollard : Will pastebin Nov 04 12:40:00 CTtpollard : http://pastebin.com/YEb61QBd Nov 04 12:40:26 cart_man: you're missing a trailing \ on the metaqt5 entry Nov 04 12:41:29 also, it's probably not wise to clone qt5 into poky Nov 04 12:41:58 Hmm ok so where should I do it? make a folder inside poky? Nov 04 12:45:00 rburton: the build issues for -next look transient, ok to merge it? Nov 04 12:46:30 chmod: cannot access '/srv/www/vhosts/autobuilder.yoctoproject.org/pub/sstate//af/sstate:glibc:core2-64-poky-linux:2.24:r0:core2-64:3:af7a0582cdaf85391d07735d19bdab2f_package.tgz.qIPf4lK5': No such file or directory Nov 04 12:46:32 whaaaat Nov 04 12:47:10 transient nfs snafu? Nov 04 12:48:59 RP: selftest is still running but its only got the expected kernel fail so far Nov 04 12:49:24 RP: yes, but remember to remove my WIP patch ;) Nov 04 12:49:29 rburton: We do see that periodically and I can only assume NFS being odd Nov 04 12:49:41 rburton: the code looks safe Nov 04 12:49:50 yeah i have the bug for that i think Nov 04 12:50:03 * RP will remember to remove WIP Nov 04 13:26:17 I am having trouble compiling gcc-4.9 Nov 04 13:27:15 I seems to use -Werror and for some odd reason the c++ compiler is used to compile some of the .c files in the project leading to warnings, again leading to errors Nov 04 13:28:14 I am not really sure where the bitbake config ends and the gcc project begins... in other words; could this be a yocto issue or should I bug the gcc boys? Nov 04 13:28:59 deva_: building gcc 4.9 with a modern distro using gcc6? Nov 04 13:29:09 rburton: Exactly Nov 04 13:29:35 what release of yocto? Nov 04 13:30:08 (we've fixed this for a few releases back in git/point releases) Nov 04 13:30:09 I followed this: https://www.compulab.co.il/workspace/mediawiki/index.php5/CL-SOM-iMX6UL:_Yocto_Linux:_Building Nov 04 13:30:50 I'm not entirely sure of the yocto version Nov 04 13:31:28 such fun Nov 04 13:32:03 dora! Nov 04 13:32:30 released october 2013 Nov 04 13:32:44 aah... that might be the problem then :-) Nov 04 13:33:17 ah, no, wrong branch Nov 04 13:33:32 fido Nov 04 13:33:45 (april 2015) Nov 04 13:34:09 yeah that doesn't support gcc6 host Nov 04 13:34:12 that's still pre gcc6 I think? Nov 04 13:34:30 To be honost I don't really understand why ubuntu ships gcc6 at all Nov 04 13:34:46 because its the current compiler Nov 04 13:34:48 works great Nov 04 13:35:09 you'll find https://www.yoctoproject.org/downloads/tools/dora15/buildtools-sdk-dora-15 useful Nov 04 13:35:26 great. thanks Nov 04 13:35:32 no, i mean https://www.yoctoproject.org/downloads/tools/fido18/buildtools-sdk-fido-18 Nov 04 13:35:48 compiler etc that is known to work with that release. Nov 04 13:35:59 ok Nov 04 13:36:00 basically an older gcc that builds the gcc Nov 04 13:36:14 I supposed installing an older version of ubuntu might fix the issue then? Nov 04 13:36:29 yeah, or that Nov 04 13:36:51 great. I'll try that Nov 04 13:36:57 buildtools is easier Nov 04 13:37:02 Better to work in a known configuration I think Nov 04 13:39:52 hi guys Nov 04 13:48:20 zeddii, can I safely add linux-yocto-4.8 as a remote on torvalds linux Nov 04 13:50:52 cart_man: sorry went AFK, any progress? Nov 04 13:51:35 CTtpollard : No I restarted the entire thing Nov 04 13:51:38 Soo will upate soon Nov 04 13:54:15 Crofton|work. shouldn't be a problem. Nov 04 14:04:01 Cttpollard: Ok soo its seems like in Order ot get Qt5 I need Openembedded-core BUT I can not have the normal Yocto and the openembedded-core one as a layer ? Nov 04 14:04:53 cart_man: qt5 is a separate layer from OE or poky, meta-qt5 Nov 04 14:05:07 you can add whatever layers you wish Nov 04 14:06:53 Yea but it sais it depends on openembedded-core -> https://layers.openembedded.org/layerindex/branch/master/layer/meta-qt5/ Nov 04 14:06:59 and when I add that it does not work Nov 04 14:07:40 ok soo that is openembedded layers ...sigh .. where do I get a list of yoctos standard layers / Nov 04 14:07:42 ? Nov 04 14:08:41 all yocto projects depend on OE core Nov 04 14:08:55 it's the build system of yocto Nov 04 14:09:21 CTtpollard : Ohhh well that makes much more sense >.< ...thouht they somehow just co-worked Nov 04 14:10:04 there are dev & ref manuals available, maybe start there for more insight Nov 04 14:12:58 CTtpollard : one more thing though.... if I added a Layer it will be included into any package I build wether it is core-image-sato or any of the other ones right ? Nov 04 14:13:46 anyone on this channel using systemd with their Yocto builds? Nov 04 14:14:15 cart_man: if you ask for a package from the layer, or a package you build has a dependency from the layer, yes Nov 04 14:14:28 if you don't specify something, bitbake will not build it Nov 04 14:15:19 ok soo how do I tell bitbake to add all of qt5 ? Nov 04 14:15:29 in local.conf file right ? Nov 04 14:15:41 Although I am not sure what I type in Nov 04 14:17:27 you can use image install append in local conf to specify extra packages on top of images yes Nov 04 14:17:47 cart_man: http://www.yoctoproject.org/docs/2.2/dev-manual/dev-manual.html Nov 04 14:21:13 Hello I'm having trouble getting the UserAdd stuff to work properly. I've got an image which includes the following recipe. But on do_rootfs it fails complaining there is no package install cadidate Nov 04 14:21:14 https://gist.github.com/strike5150/e7da196d5140a5d0f884a5364b2a8b25 Nov 04 14:25:53 Strike5150: no need to add FILES because you don't actually create those files, remove INHIBIT_PACKAGE_DEBUG_SPLIT as that does nothing useful here, and add ALLOW_EMPTY_${PN} = "1" to actually fix your problem Nov 04 14:26:14 Strike5150: (as you didn't create /home/newuser the package has no files in, and empty packages are not created unless you say otherwise) Nov 04 14:32:25 Ahh thank you :D makes sense Nov 04 14:35:17 so i got a bug report about pseudo because of systems where "python" is "python3". Nov 04 14:35:40 RP: Any objections to a patch to configure which tries to find a likely Python version, and/or lets the user specify it? Nov 04 14:36:12 I figure bitbake's the largest user by far, but apparently something else recently picked up pseudo, so. Nov 04 14:37:16 seebs: being able to configure it is fine with me Nov 04 14:38:07 Proposed plan: If not specified, search $PATH for python, check to see whether it seems to be v2-ish, then try v2, and otherwise explode in rage and fury, deleting every file the user ever loved or cared about. Nov 04 14:38:15 ... wait, i'm not a minecraft modder. I'll omit the deleting files part. Nov 04 14:38:32 seebs: sounds good :) Nov 04 14:38:40 seebs: any chance we can make it just need python3? Nov 04 14:38:49 seebs: we're trying to move to a py3 only world Nov 04 14:39:45 I've been hesitant about the upgrade, but a lot of that is historical from WR, where we got stuck with RHEL4 for I think four years longer than we should have been. Nov 04 14:39:54 I could probably just change it to python3 and hope for the best. Nov 04 14:40:17 I'd have to fix the code, though. Yay for stable languages that don't have a ton of breaking changes all the time. Nov 04 14:40:26 I think there is something to try to guess python version when creating the self-extracting stuff Nov 04 14:41:10 And of course, the machine I'm on right now hasn't even got a python3. AT ALL. Nov 04 14:41:20 (Oldish Mac OS X.) Nov 04 14:42:59 https://github.com/openembedded/openembedded-core/blob/master/meta/files/toolchain-shar-relocate.sh Nov 04 14:43:31 seebs: line 27 Nov 04 14:43:56 Yeah, although that doesn't seem version-aware, so I assume the thing it calls will run with either? Nov 04 14:43:58 (if that helps) Nov 04 14:44:20 yes, what it calls runs with either Nov 04 14:44:46 I'm not as confident of my ability to make the code run with either, although I suppose I could look at that program and see what it does. It might be easy enough. Nov 04 14:57:29 seebs: can you point me to the bug report? Someone mentioned a similar issue to me. Nov 04 14:57:46 I don't know of a bug report, I just got an email from a buildroot user. Nov 04 14:58:08 seebs: ah---probably the same Nov 04 14:58:54 guy in my office was fighting with python3+pseudo+buildroot so I'm guessing he filed a bug ;-) Nov 04 14:58:56 thx Nov 04 14:59:41 Gaël? Nov 04 15:00:31 yep---that's him! Nov 04 15:23:02 Can somebody point me in the right direction to add my own /etc/network/interfaces. I've tried using a bbappend to init-ifupdown, but the interfaces file itself always ends up being the original not my overlayed version Nov 04 15:28:08 I am using the following bbappend to try and change the interfaces file, but its failing to overlay it. Nov 04 15:28:09 #Appends ~/poky/meta/recipes-core/init-ifupdown SRC_URI = "file://interfaces" do_install_append() { # Install file only if it has contents install -d ${D}${sysconfdir}/network install -m 0644 ${S}/interfaces ${D}${sysconfdir}/network/interfaces } do_install_append_qemuall () { } Nov 04 15:28:15 https://gist.github.com/strike5150/e7da196d5140a5d0f884a5364b2a8b25 Nov 04 15:55:59 Hi, Is this possible to check what needs to be rebuild after change in one of recipes and what will be used from cache with bitbake command? Nov 04 16:00:56 TomK: try bitbake-whatchanged Nov 04 16:05:40 TIL Nov 04 16:05:43 that's cool Nov 04 16:12:11 khem: It shows a lot more packages than quick summary with dry-run (this : Attempted 5717 tasks of which 5676 didn't need to be rerun). Does dry run don't check if package has action compile then it also needs to update packages that depens on that? Nov 04 16:19:03 it should tell you all that bitbake will do Nov 04 16:19:15 many a times its only quick tasks Nov 04 16:22:34 I am using the following bbappend to try and change the interfaces file, but its failing to overlay it. Nov 04 16:22:42 https://gist.github.com/strike5150/e7da196d5140a5d0f884a5364b2a8b25 Nov 04 16:23:56 Strike5150: you need to let bitbake know about the new FILEPATH Nov 04 16:24:26 RP: i want to get the checksum of a specific variable/function which isn't a task, obeying the basehash whitelist. is there anyway to do that without basically duplicating parts of _build_data() and generate_dependencies()/ Nov 04 16:24:36 RP: seems a bit odd, if so Nov 04 16:25:21 Strike5150: something like FILESEXTRAPATHS_prepend := "${THISDIR}/files:" provided you stored the interfaces file under files/ dir Nov 04 16:25:23 khem: I'm not clear on what you mean by that, are you saying its not using the recipe because it thinks it shouldn't take precedence? Nov 04 16:25:59 khem: Ok I think your saying its not adding it to the files of the parent recipe Nov 04 16:26:28 Hi all. I have an issue with incorrect logrotate permissions in my poky image. In the rootfs tar.gz they are "-rw-rw-r-- 1 root root 896 Jun 23 17:05 logrotate.conf" but in the ipk they are "-rw-r--r-- 115/65534 896 2016-10-10 21:29 ./etc/logrotate.conf".. Any ideas why they might be getting the extra g+w when tarred? Nov 04 16:26:37 additional info: building the same image in two different locations. In one the permissions in the rootfs tarball are the correct 644, in the other they come out as 664, so it seems like an environment issue Nov 04 16:33:29 khem: Thanks for help. I think that I will try to figure out the rest Nov 04 16:45:43 khem: for the FILESEXTRAPATHS_prepend := "${THISDIR}/files:" do you mean init-ifupdown/init-ifupdown-1.0/interfaces. Or init-ifupdown/files/interfaces Nov 04 17:30:14 Hi! Has anyone worked with systemd-networkd? I'm trying to enable it by adding an entry to DISTRO_FEATURES_append for "networkd", but when I examine the result in my rootfs I'm missing the networkctl binary and the systemd-networkd.service file Nov 04 17:32:04 dscully: pretty sure networkd is a packageconfig in the systemd recipe, not a distro feature Nov 04 17:38:48 Ah, thank you! That lead to me the right spot in the docs Nov 04 17:41:36 np **** BEGIN LOGGING AT Fri Nov 04 18:26:21 2016 Nov 04 18:40:12 As I understand adding: FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}-${PV}:" to a bbappend should trigger yocto to override any files in the ${PN}-${PV} directory of the same name as the original recipe Nov 04 18:40:20 Does that look right? Nov 04 18:42:24 Does that include subdirectories or do they need to be added explicitly Nov 04 18:43:37 I can't think of anything that implicitly searches subdirectories in bitbake. Nov 04 18:43:45 ok Nov 04 18:43:48 thx :D Nov 04 18:44:50 I suppose that would be silly, should have thought of that :P **** ENDING LOGGING AT Sat Nov 05 03:00:00 2016