**** BEGIN LOGGING AT Mon Oct 23 03:00:01 2017 Oct 23 06:35:17 Hi, does anyone know what's wrong with my generated core-image-minimal? Basically udevadm info isn't getting all the properties of an image overlay in QEMU. Here are the differences between running the command in core-image-minimal and a regular debian image with the same image overlay https://imgur.com/a/dlqn6. Any idea how I can fix this? Oct 23 07:51:14 hello, how do I have a core-image based image that uses rpm as packaging system, but does not include dnf as a tool? I am planning to use swupdate instead of dnf to install packages, so how do I remove the dnf and libdnf packages from final image? Oct 23 07:51:22 This is on Yocto Pyro Oct 23 07:52:39 Getting a systemd failure for dnf-makecache.service is kind of annoying and inappropriate, so I would like to get rid of dnf alltogether, but not sure how it really gets included Oct 23 08:01:51 eduardas_m: you want to use swupdate to install rpm ? Oct 23 08:03:58 nayfe: both for fully overwriting rootfs (major updates) and installing (or updating) rpms (minor updates), yes Oct 23 08:04:18 not sure how much that makes sense, but the ability for swupdate to do that is there Oct 23 08:04:39 eduardas_m: for my inderstanding, swupdate is not meant to handle packages Oct 23 08:06:20 eduardas_m: https://groups.google.com/forum/#!topic/swupdate/5LCXDhK-6U0 Oct 23 08:08:21 eduardas_m: may I ask why you don't want dnl ? Oct 23 08:09:22 nayfe: I just did not see a reason why an rpm could not be added to a software collection and installed via script added to swupdate package Oct 23 08:09:59 assuming your image inherits core-image then remove package-management from DISTRO_FEATURES to prevent the package manager from being installed Oct 23 08:10:12 nayfe: it kind of does not make sense to me to have two update mechanisms: both swupdate and dnf Oct 23 08:10:48 eduardas_m: well image update and pkg update is really different kind of things :) Oct 23 08:11:02 joshuagl: I want both the packages and rpm tool, I just don't need dnf specifically Oct 23 08:12:03 because there won't be a cloud-accessible rpm repository to maintain if I distribute the rpms via swupdate software collections Oct 23 08:12:35 my resoning might be severly flawed, but that is my trail of thought Oct 23 08:13:13 nayfe: so how would you go about it if you need both? Oct 23 08:13:22 "and installed via script added to swupdate package" => for me this script should be dnl ? Oct 23 08:13:44 eduardas_m: remove package-management from IMAGE_FEATURES and add rpm to IMAGE_INSTALL ? Oct 23 08:14:27 joshuagl: ok, will try that... thank you Oct 23 08:15:07 the package-management feature maps to the ROOTFS_PKGMANAGE variable, in rootfs_rpm that's = "rpm dnf" :-) Oct 23 08:16:44 nayfe: if I leave dnf after all, how do I appropriately prevent dnf-makecache.service from failing... it fails because there are no external repos Oct 23 08:17:35 Error: There are no enabled repos. Oct 23 08:17:49 dnf-makecache.service: Main process exited, code=exited, status=1/FAILURE Oct 23 08:17:58 mask or remove the service? Oct 23 08:19:11 https://fedoramagazine.org/systemd-masking-units/ Oct 23 08:20:27 joshuagl: yeah, my first thought was to simply disable it... was not sure if that was entirely appropriate though... thanks for your opinion Oct 23 08:20:58 np Oct 23 08:49:58 the dnf_git.bb recipe has some systemd units enabled by default. Is it approprate to disable them via a bbappend? if so, what would be the proper mechanism to do that? the point is that I need those units disabled the first time I boot the image Oct 23 09:48:11 hum strange thing, got two identical DTS, making fitimage with 2 configurations with those 2 DTS, same kernel, one is booting, the other hangs, anything am I missing ? Oct 23 09:53:03 nayfe: just out of curiosity... documentation says "kernel-fitimage class provides support to pack zImages"... what does that even mean? was unable to google anything meaningful on the topic Oct 23 09:54:54 I'm not using kernel-fitimage, but it means it will package zImage with DTB files and optional ramdisk into a FIT image Oct 23 09:59:41 nayfe: this is interesting... haven't heard about those before... any specific motivation for using a FIT image? Oct 23 10:00:28 to ease and secure updates Oct 23 10:28:50 hi.. how to call in image recipe a python function from variable expansion like: python my_set_cflags_release_switch() { return "-DTEST" } TARGET_CFLAGS += "${@my_set_cflags_release_switch()} Oct 23 10:29:52 i get "NameError: name my_set_cflags_release_switch' is not defined." Oct 23 10:30:10 it works using @oe.utils.ifelse Oct 23 10:48:29 or, how to add own lib/*.py scripts to path used by bitbake? Oct 23 11:03:39 Is there some cool trick to show only the vars i needs with bitbake -e? something like bitbake -e -filter "IMAGE_FSTYPES" Oct 23 11:05:19 grep ? :-) Oct 23 11:05:26 I'm writing my first recipe and used devtool add as a starting point. The recipe is for a python module. According to the commets it could not find map standard stuff like 're' and 'collections' to a recipe. Is this common? Oct 23 11:25:59 I still couldn't fix my wic dependency problem :( Oct 23 11:28:20 styler2go_: what is the exact problem? could you elaborate? I also use wic... Oct 23 11:29:07 Sure! I have the problem that i want to use wich but wic executes simultanously with do_hddimg, do_bootimg etc. and thus it fails because it cannot find the file Oct 23 11:29:21 i always have to execute bitbake 3 times until all image types are created Oct 23 11:29:56 ping rburton Oct 23 11:30:19 hi Oct 23 11:31:16 hi rburton: do you know how i can add python scripts to my layer in lib/ so they can be found by bitbake and to be used inside ie local.conf? Oct 23 11:31:41 like the one from poky in meta/lib/oe Oct 23 11:32:03 styler2go_: I assume you have some final image that has IMAGE_FSTYPES = "wic"? And you want to have some components in deploy directory to use with wic before final image creation? Oct 23 11:33:11 eduardas_m, well, if i don't have the iso file it fails with an error: exec_cmd: cp /mnt/upboard/build-recovery/tmp/work/up_board-poky-linux/recovery-image/1.0-r0/recovery-image-1.0/hddimg/EFI/BOOT/* /mnt/upboard/build-recovery/tmp/deploy/images/up-board/recovery-image-up-board-20171023112547/build/hdd/boot/EFI/BOOT returned '1' instead of 0 Oct 23 11:33:44 btooth: yes, put them in lib/ Oct 23 11:34:19 eduardas_m, i somehow need wic to wait until do_bootimg is done Oct 23 11:36:08 eduardas_m, i've now set do_image_wic[depends] += "xxx:do_bootimg" and now it works Oct 23 11:36:34 styler2go: I am not exactly sure about your usecase, but can you not separate out any artifact creation you try to do during do_bootimg into separate bitbake recipe and then add a DEPENDS to final image recipe..... that is how I go about artifact creation Oct 23 11:37:09 Hmm i don't do anything special.. I just build an ext4 for my sdcard Oct 23 11:37:35 And now i found out about wic and i would like to use that, but wic executes before hddimg is done Oct 23 11:37:35 styler2go: you have a separate ROOTFS for recovery? Oct 23 11:38:25 bitbakes creates everything on it's own Oct 23 11:38:27 rburton: in my local.conf i do:_ TARGET_CFLAGS += "${@testme.testmefunc()}" and in my lib/ folder i have testme.py with def testmefunc() --> I get: NameError: name 'testme' is not defined Oct 23 11:38:30 it creates the rootfs Oct 23 11:39:21 styler2go: what does hddimg do in your case? why do you need it? Oct 23 11:39:30 wic needs it Oct 23 11:39:45 styler2go: not necessarily, depends on usecase Oct 23 11:39:47 i am using predefined mkefidisk if that's important Oct 23 11:39:51 I use wic without it Oct 23 11:40:20 In my case it produces the error above until the hddimg exist Oct 23 11:40:23 btooth: did you import testme? Oct 23 11:40:59 styler2go: so you are trying to work with a system that uses EFI? Oct 23 11:41:03 x86? Oct 23 11:41:05 yes Oct 23 11:41:26 rburton: where should i do the import? i can not do inside local.conf right= Oct 23 11:41:31 if i add the line above it works now.. but it still creates 4 files: ext4, hddimg, iso and finally the wic file Oct 23 11:41:42 styler2go: I am sorry. If that is the case, I won't be able to help you. My usecase is an ARM system with u-boot. Oct 23 11:41:43 https://p.styler2go.de/4490738 Oct 23 11:42:12 eduardas_m, no problem! but thanks for your help anyway Oct 23 11:50:18 btooth: right. probably easier to make a class which has a function in, then you don't even need to drop pure py into lib/ Oct 23 11:51:39 rburton: okay i will try this, but i am wondering how poky sets path for bitbake to find scripts in lib/oe.. so would be good to know.. if you have any idea please let me know Oct 23 11:52:03 btooth: python include path includes lib/ in each layer Oct 23 11:52:25 you can probably do the import inside the ${@ Oct 23 11:52:37 rburton: is there a way to doublecheck this? Oct 23 11:52:48 okay good idea Oct 23 11:52:53 i will try Oct 23 11:53:03 meta/classes/base.bbclass: sys.path[0:0] = [os.path.join(dir, "lib") for dir in bbpath] Oct 23 11:53:19 there you go ^ Oct 23 11:55:54 rburton: sorry, but how to call import: ${@import testme testme.testmefunc()} ? Oct 23 11:56:06 put a ; in between the import and the call Oct 23 11:56:12 its just standard python Oct 23 11:56:23 RP1: 'sstate.bbclass: sstate_hardcode_path(): fix for multilib' <-- leaving that for you to review :) Oct 23 11:58:50 rburton: i get "SyntaxError: invalid syntax" with: TARGET_CFLAGS += "${@import testme; testme.testmefunc()}" Oct 23 11:59:14 probably limitations on what the inline py can do Oct 23 11:59:19 as i said, use a class to define a function Oct 23 12:12:03 rburton: thank you, the paste to base.bbclass helped me, i can set OE_IMPORTS += "testme" in local.conf and now my function is executed Oct 23 12:31:21 rburton: ok, not now though :) Oct 23 14:10:52 Does anyone know if there is a registration fee for OEDAM if you are already going to the ELC? Oct 23 14:15:22 OEDEM/OEDAM are free Oct 23 14:15:31 OEDEM was yesterday in Prague Oct 23 14:25:05 hi everyone Oct 23 14:25:44 hmm, since today i have a weird problem. when i try to run bitbake it stalls at 99% parsing recipes and when i try to stop it the bitbake server is not responding Oct 23 14:26:04 i've reduced the test project to just a vanilla poky:pyro Oct 23 14:26:47 sounds like you need to call ghost^W^W^W^W^W strace Oct 23 14:27:00 joshuagl: Ya, I was specifically wondering about the North America conference. Thanks! Oct 23 14:27:12 JPEWhacker: np, just wanted to make sure :-) Oct 23 14:30:40 huh, killing the process, manually killing the bitbake server, removing the lock file and then starting it again apparently works Oct 23 14:30:43 ....what Oct 23 15:22:46 hello guys. I have a problem using "repo". I have several git repos that are at different levels in the main project. I want to archive them all aloong with their position from the root directory of the project. to do so I ve used the following command: repo forall -c 'git archive --format=tar --prefix=$REPO_PATH/ HEAD > P.tar; tar --concatenate --file='`pwd`/all.tar' P.tar; rm P.tar'. By the way it seems to add only the first level git Oct 23 15:22:46 repos. Does anibody know how to solve the issue ? Oct 23 15:27:31 Is it possible to merge sstate and download dirs from two machines? Would it work with a plain rsync? Oct 23 15:38:26 sveinse: yes Oct 23 16:06:28 hi guys Oct 23 16:06:57 do we have some maintainers here? Oct 23 16:07:00 http://lists.openembedded.org/pipermail/openembedded-core/2017-October/143537.html Oct 23 16:07:11 any blockers on getting this one to the pyro branch? Oct 23 18:17:56 rburton, maybe? Oct 23 18:18:07 Hello all! How do I quickly find out who or what sets an environmental variable with bitbake? the -e only shows the values Oct 23 18:18:09 http://lists.openembedded.org/pipermail/openembedded-core/2017-October/143537.html Oct 23 18:18:43 Adam_: any reasonably modern bitbake will show what recipe assigned the values in -e Oct 23 18:21:16 rburton, -e output is long. And if I grep for the variable of my interest, it only shows the final value.. Oct 23 18:21:37 Adam_: pipe through less and then search using / and scroll up Oct 23 18:21:51 grep is no good as it will *only* show the assignment and not the lines before it Oct 23 18:22:28 rburton, ah Ok. I was hoping to find a more convenient mechanism, but that will do! Oct 23 18:22:48 Adam_: patches welcome for a "just this variable" mode Oct 23 18:23:09 rburton, I will see if I can put something together ;). Thanks for the opportunity Oct 23 18:37:28 Can anyone here please tell me what is the status of the external-linaro-toolchain recipe (and associated tcmode conf file) in the meta-linaro-toolchain layer? As far as I can tell it's broken in a number of ways (e.g. claims to provide glibc, but doesn't actually ship a TCLIBC configuration file). Oct 23 18:42:46 laplante, TI's SDK use external toolchain and it is building fine for me - http://arago-project.org/git/?p=meta-arago.git;a=blob;f=meta-arago-distro/conf/distro/include/toolchain-linaro.inc;h=fd0397d2827d0a6b6dcef96d3406e6020eb4b31b;hb=HEAD Oct 23 18:45:05 Adam_: Thanks! As luck would have it, I've already incorporated their toolchain inc file into our project, which does seem to get at least target builds working. Should also mention I'm using the 7.1.1 release of Linaro (gcc 7). Oct 23 18:46:08 Adam_: The main thing broken for me right now is lack of nativesdk support. Oct 23 18:46:44 laplante: np! -c populate_sdk doesn't work for you? Oct 23 18:48:36 Adam_: Unfortunately not. During the preamble to my image builds, Yocto complains: "NOTE: preferred version linaro-7.1% of nativesdk-libgcc not available (for item nativesdk-libgcc) \nNOTE: versions of nativesdk-libgcc available: 5.4.0 6.2.0 linaro-4.9 linaro-5.2 linaro-5.3 linaro-6.2" Oct 23 18:49:09 Adam_: Which makes sense I think - I don't see anything in the external Linaro stuff to indicate that they've inherited from the nativesdk class. Oct 23 18:49:57 laplante: Ah I see. I am still stuck with GCC 6.2. Sounds like linaro-7.1 is doesn't supoprt -native build.. Oct 23 18:51:34 Adam_: That's unfortunate :(. Do you use any kind of TCLIBC conf file to choose external-linaro-toolchain as glibc provider? Or do you build glibc and friends from the built-in recipes (i.e. you just use the external toolchain for the binaries, and use those to generate the actual sysroot)? Oct 23 18:54:48 laplante: I use the tool-chain provided by TI (which appears to be Linaro's). It's packaged binary. No building from scratch. Oct 23 18:56:29 laplante: but, I am trying to build another set of toolchain using the external toolchain (Cross-canadian..?) Oct 23 18:58:25 Adam_: I think I'll have to clone the TI stuff and see how they do it. The external Linaro support in meta-linaro seems incomplete... which makes me wonder why Yocto bothers including it. Oct 23 18:59:08 Adam_: Ah, yes. AFAIK cross-canadian is a sort of synonym for native SDK. So, your "SDK machine" is the third machine (along with host and target), making it a "Canadian cross". Oct 23 18:59:20 laplante: Take a look at the TI's setup. Keep in mind it's not as bleeding edge as gcc-7 Oct 23 18:59:48 Adam_: Will do! Thanks so much for your help. Oct 23 18:59:58 laplante: I see. Canadian's a culturally diverse I heard... Oct 23 19:00:02 np! Oct 23 19:01:32 Adam_: lol indeed. Oct 23 19:12:08 Adam_: Ah-HA! meta-arago indeed messes a ton with the external Linaro stuff to get it to work: http://arago-project.org/git/?p=meta-arago.git;a=tree;f=meta-arago-extras/recipes-core/meta;h=f0e329146d1630713b8bf6cdf71fa827b78e2fb9;hb=HEAD. The meta-arago-extras layer itself defines cross-canadian support for external Linaro. Oct 23 19:12:48 I don't know if this is the right forum or not, but our local builds have been failing to pull meta-java and meta-mono for several days now. Does anyone have an idea of what may be going on? Oct 23 19:14:45 jhatch2008: Where is it trying to pull from? e.g. github, openembedded.org Oct 23 19:15:12 laplante: haha ya, they definitely put a lot of work on it to get their ext-toolchain to work Oct 23 19:15:37 laplante: in other words, you can get yours working too... with some effort :) Oct 23 19:15:51 Adam_: That's pretty annoying... those changes should just be in meta-linaro itself. I'll write to the mailing list to ask why... Oct 23 19:16:38 laplante: Sounds good. Btw, where is meta-linaro mailinglist? Oct 23 19:16:45 Adam_: Yup, I've spent almost a week hand-tweaking it, and so far have only been able to get it to "kind-of" work. Oct 23 19:17:24 laplante: Sounds like it's about time you raise an issue ;) Oct 23 19:17:26 Adam_: Looks like openembedded@lists.linaro.org Oct 23 19:17:32 laplante: ah ok Oct 23 19:17:33 Trying to pull from 'git://git.yoctoproject.org/meta-java' Oct 23 19:18:12 Adam_: Certainly will. Google is full of people with issues like those I encountered, but barely any solutions. Turns out those solutions were hiding in meta-arago, of all places. Oct 23 19:19:20 laplante: Trying to pull from 'git://git.yoctoproject.org/meta-java' Oct 23 19:19:41 laplante: Glad you found it. TI did a good job on getting their external toolchain together.. although I am not 100% sure if that was necessary to begin with.. Oct 23 19:21:45 laplante: Keep getting the following error: fatal: unable to connect to git.yoctoproject.org: git.yoctoproject.org[0: 140.211.169.56]: errno=Connection timed out Oct 23 19:23:29 jhatch2008: What happens when you just try to ping git.yoctoproject.org? Oct 23 19:24:56 laplante: I'll double check but that was timing out on our Jenkins server. I have not seen the problem from my PC, but other people have been seeing it here. Give me a minute to check the server. Oct 23 19:28:17 laplante: ping fails, though DNS is working: jenkins@jenkinsdm1:~$ ping git.yoctoproject.org PING git.yoctoproject.org (140.211.169.56) 56(84) bytes of data. ^C --- git.yoctoproject.org ping statistics --- 5 packets transmitted, 0 received, 100% packet loss, time 4032ms Oct 23 19:28:34 laplante: I get the same thing from my PC. Oct 23 19:29:07 isn't it common to block ICMP? So ping might not be a reliable test if the service is present Oct 23 19:31:53 jhatch2008: I'm able to clone the meta-java repo from git.yoctoproject.org, btw Oct 23 19:31:55 sveinse: Good point Oct 23 19:32:06 sveinse: I agree, however telnet to yoctoproject.org on 443 just hangs also Oct 23 19:32:25 jhatch2008: Can you wget the homepage? Oct 23 19:32:47 halstead: ping Oct 23 19:33:31 LetoThe2nd: pong Oct 23 19:33:43 halstead: see above Oct 23 19:33:50 if it's www.yoctoproject.org you're referring to, then that service is up from here Oct 23 19:33:55 laplante: no I can't wget yoctoproject.org either https or http Oct 23 19:33:56 One sec Oct 23 19:34:41 jhatch2008, laplante: wget works here Oct 23 19:35:13 svalan, I'm here to help troubleshoot. Oct 23 19:35:28 jhatch2008: Sounds like a firewall/network issue then Oct 23 19:35:33 sviense: Could you do me one small favor an let me know if it's the same IP that you're going to: 140.211.169.56... Oct 23 19:35:42 jhatch2008: hang on Oct 23 19:36:23 sveinse, Can you test a unencrypted page as well like, http://www.yoctoproject.org/docs/2.3.2/bitbake-user-manual/bitbake-user-manual.html Oct 23 19:36:28 jhatch2008: That's the IP address I'm getting Oct 23 19:36:45 jhatch2008: yes, the same IP and it does reply Oct 23 19:37:12 laplante: Thanks, I think it is local to us. I'll be visiting our IT. Oct 23 19:37:18 halstead: yep, is working here too Oct 23 19:38:15 halstead: thought you are the guy to look into this.. .and sorry for the naked ping, mind is being rather slow by now Oct 23 19:38:55 LetoThe2nd: ha, I just assumed that was a bit you two had with each other :) Oct 23 19:39:12 I am :) LetoThe2nd, no problem. I just didn't have the full history in that client. Oct 23 19:39:25 have you tracert'ed it to see how far off this thing is blocked? ICMP is apparently ok with this server Oct 23 19:39:48 laplante: nah. but when i read something that i can't help with but think i know the person who cen, then i... well, ping him/her Oct 23 19:40:10 LetoThe2nd: perhaps he has a ICMP-block installed? :P Oct 23 19:40:37 sveinse: to be honest, i don't have the slightest clue at all. Oct 23 19:40:56 LetoThe2nd: (last one was a terrible joke) Oct 23 19:41:26 sveinse: you see how clueless i am when it comes to networking, i don't even get the jokes. Oct 23 19:41:34 jhatch2008, So both "git clone https://git.yoctoproject.org/git/meta-java" and "git clone git://git.yoctoproject.org/meta-java" timeout for you? Oct 23 19:42:51 halstead: if relevant and for reference, both works from here Oct 23 19:44:27 Thanks sveinse. That is helpful. I'm not having trouble from multiple DCs. Maybe there is a bad route for jhatch2008? Oct 23 19:45:38 i can offer checks from prrague and munich if needed Oct 23 19:46:09 LetoThe2nd, Sure. Thank you. If my pong back was impolite I apologize. It was a common thing in another community. Oct 23 19:55:15 clones from munich work fine too Oct 23 19:58:40 and prague is fine too Oct 23 20:04:55 halstead: Yes, both. I believe that it is either local or we are being blacklisted somehow. Talking with our IT folks. I'll be back if we determine it's not local. Oct 23 20:05:24 jhatch2008, Okay. I can give you an alternate address if you'd like to test. Oct 24 00:38:07 RP1, hi, any blocker for http://lists.openembedded.org/pipermail/openembedded-core/2017-October/143537.html ? Oct 24 00:45:56 otavio, maybe ? Oct 24 00:54:14 Is there a yocto analog to the rootfs overlay concept in buildroot? Oct 24 01:13:14 tront, RP1 is not blocking that patch. Its in a patch series I am trying to validate prior to asking for a merge request to pyro Oct 24 01:15:52 thanks for the reply Oct 24 01:16:15 any rough idea on how much time it would take? Oct 24 01:21:55 not sure. I am working throw build errors so a few more days. Oct 24 01:23:46 any chance you could separate that one from the rest that are throwing build errors now and ask for a merge request solely for it? :P Oct 24 01:24:13 we're really needing this in pyro for a while to unblock some stuff Oct 24 01:27:07 you can ask on the mailing list for RP to merge. The errors I am seeing may not be related to the changes I have queue rather the AB build env / hosts. Oct 24 01:27:57 pyro proper now failing w/o my patch series Oct 24 01:29:54 armpit, we have asked for this to get merged in the pyro branch: http://lists.openembedded.org/pipermail/openembedded-core/2017-October/143537.html Oct 24 01:36:44 I realize that. I suspect RP is waiting on the green light from me while he focuses on the Rocko release. You are welcome to re-ping. Oct 24 01:40:52 armpit, reping in that thread or should somehow the same commit be sent to the mailing list targeted at the pyro branch? Oct 24 01:41:26 just asking what the procedure usually is Oct 24 01:42:19 just re-ping to your last ping is the norm Oct 24 01:43:09 RP would be the one to make the call in this case. Oct 24 01:44:10 thanks Oct 24 02:05:12 Hi there, I'm trying to debug a Yocto daisy build that is failing in create_sdk_files. Error: movefile failed to move /path/to/sdk/image/var/lib/opkg/status to path/to/sysroots/i686-clsdk-linux/var/lib/opkg Oct 24 02:05:31 OSError(21, 'Is a directory')) Oct 24 02:06:11 and ideas? I've found references to this problem online, but with no one has posted and solution. Oct 24 02:08:15 OS2.1 has been retired since 1997. You might be stuck in a historical trap. Oct 24 02:10:22 Hi there, I'm trying to debug a Yocto daisy build that is failing in create_sdk_files. Error: movefile failed to move /path/to/sdk/image/var/lib/opkg/status to path/to/sysroots/i686-clsdk-linux/var/lib/opkg Oct 24 02:10:31 OSError(21, 'Is a directory')) Oct 24 02:10:59 has anyone seen something like this before? Oct 24 02:20:20 I'd try to clean the build and bitbake again Oct 24 02:21:57 Try and clean the recipe that is failing first, then try a cleanall on it, and progress from there before potentially starting a fresh build. If it is trying to move files to an already existent directory it sounds like a previous task only partially completed Oct 24 02:22:32 at least that would be my approach, but I can't speak to your specific problem Oct 24 02:28:38 thanks for the tips bodangly. I am trying similar things with no success so far. Oct 24 02:30:51 do you get the error on a fresh build? Oct 24 02:31:11 what recipe is it you are getting stuck on? Could you pastebin the full output? Oct 24 02:31:14 Yes, from a completely clean workspace. Oct 24 02:32:03 I added a compile time option to qt5 ( add sqlite support). I have reverted that change and I am trying the build again to make sure it is not a problem with my environment. Oct 24 02:32:55 If that build is successful, then my addition (PACKAGECONFIG_append = " sql-sqlite") is the problem. Oct 24 02:33:19 Then, I can debug it from there. Oct 24 02:34:15 are you doing just PACKAGECONFIG_append or PACKAGECONFIG_append_pn-qt5 (or similar)? Oct 24 02:34:59 ie are you modifying the recipe itself? Oct 24 02:35:33 I wonder if something else is also trying to copy sql-sqlite into sysroot first and isn't playing nice Oct 24 02:35:56 I know if things conflict / already exist there it causes problems Oct 24 02:36:33 I've had that problem before with a recipe that used perl Oct 24 02:36:39 I am doing PACKGECONFIG_append. Oct 24 02:38:27 within the qt5 recipe itself? Oct 24 02:38:28 I have a qtbase_5.2.1.bbappend in my custom meta-layer. That is where I am putting the PACKAGECONFIG_append. Oct 24 02:38:33 oh ok Oct 24 02:39:20 maybe try this: PACKAGECONFIG_append_pn-qtbase = " sql-sqlite" Oct 24 02:39:22 in local.conf Oct 24 02:39:55 ah. OK. I will try that next. I appreciate the help. Oct 24 02:40:05 no problem Oct 24 02:41:10 just paying it forward, this channel and IRC in general has been good to me, I'm a consultant and often end up asking questions that get solved and directly make me money :) Oct 24 02:41:28 (me too) :) Oct 24 02:41:35 lol nice Oct 24 02:41:54 its amazing how companies can sit on a problem for weeks, and no one thinks to just ask the community Oct 24 02:42:08 its like they treat open source as commercial software without a support license Oct 24 02:42:26 indeed. Oct 24 02:42:26 and ignore that a community exists Oct 24 02:42:51 I can't complain though, their ignorance puts money in my pocket, which I am sure you experience as well :) Oct 24 02:43:06 I understand what you are saying. Oct 24 02:43:39 I even usually tell them "I asked in IRC" and they look at me like I did some black magic Oct 24 02:43:57 http://www.lmgtfy.com/ Oct 24 02:44:08 :) Oct 24 02:44:48 haha yeah exactly Oct 24 02:47:05 I found a similar config setting in my local.conf, so I've dealt with this before. I think that might be the answer. I've reset the build and am starting again. Oct 24 02:47:14 cool, good luck Oct 24 02:47:31 I should know in about 3 hours since I'm starting the build from scratch. Oct 24 02:47:34 thanks! **** ENDING LOGGING AT Tue Oct 24 03:00:00 2017