**** BEGIN LOGGING AT Tue Feb 05 03:00:01 2019 Feb 05 04:43:03 New news from stackoverflow: Can't log in beaglebone as root [closed] Feb 05 07:50:03 Morniiing! Feb 05 07:51:45 Which is the proper way to extract Qt5 SDK (SKD that allows cross-compilation via QtCreator for my TARGET) in Yocto Thud? Feb 05 08:02:39 Is bitbake meta-toolchain-qt5 enough or should I follow further (previous) steps? Feb 05 08:22:38 malanecora: bitbake meta-toolchain-qt5 Feb 05 08:23:23 mckoan: That's all? I've just found this on stackoverflow https://stackoverflow.com/questions/40258161/how-do-i-get-qt-sdk-configured-properly-with-yocto-project/40375372 Feb 05 08:30:43 malanecora: depends on your needs, the basic Qt5 SDK is meta-toolchain-qt5 Feb 05 08:42:23 mckoan: Ok, ty! I think that it'll be enough Feb 05 10:14:09 New news from stackoverflow: how to apply patch uboot: Yocto project Feb 05 10:21:04 RP: yep, python hardcodes /usr/lib in a number of places, I fixed some, but apparently not all Feb 05 10:29:01 kanavin: are they arch independent pieces? Feb 05 10:30:06 RP: as far as I remember it's a mix of .py and .so Feb 05 10:30:37 kanavin: right, .py would be arch independent, .so would not Feb 05 11:18:59 Getting "ERROR: Feature 'xcomposite-glx' was enabled, but the pre-condition 'features.wayland-server && features.opengl && !features.opengles2 && libs.xcomposite && libs.glx' failed." when running bitbake meta-toolchain-qt5 Feb 05 12:39:21 RP: the current python recipe carries an awful patch that forces ${libdir} everywhere where python hardcodes /usr/lib. I'm going to basically drop that, as /usr/lib bits are either arch independent, or go into an arch-specific subdirectory in /usr/lib. Feb 05 12:39:25 RP: the patch is https://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-devtools/python/python3/python-3.3-multilib.patch Feb 05 12:40:13 kanavin: I'd be happy if we could make it work more neatly/properly with multilib Feb 05 13:25:58 hello. i have in glibc-initial.inc "deltask do_build" line, but when try to build (MACHINE=aarch64-sdk-generic bitbake -v nativesdk-glibc-initial), get error Task do_build does not exist for target nativesdk-glibc-initial. What can be the problem? Feb 05 13:29:28 inisider: specify a task which exists with the -c option Feb 05 13:30:03 inisider: you almost never want to build glibc-initial directly Feb 05 13:36:44 Hi all, do you have any recipes example based on meson build system ? :D Feb 05 13:41:29 several in oe-core Feb 05 13:41:37 $ bitbake-layers show-recipes -i meson Feb 05 13:43:04 thx Feb 05 13:43:18 kanavin: I suspect you'd approve of http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/t222&id=075820c4d768aee33badfcfd20930f0de76f08b6 :) Feb 05 13:43:50 vmeson: ^^^ I suspect you and Mark may disagree with this one though Feb 05 13:45:05 RP: I approve, totally :) (and especially rburton would like this, as he's been working on a crazy project to make them kind of work as proper src.rpm) Feb 05 13:54:29 kanavin: deleting archiver.bbclass is tempting too... Feb 05 14:51:52 RP: We use archiver.bbclass extensively, but it does seem over complicated (perhaps it is *too* configurable?) Feb 05 14:52:22 kanavin: oh i have not been working on a project to make them work Feb 05 14:52:27 kanavin: you take that back! :) Feb 05 14:52:40 i *did* and it was formative in my incredibly strong opinions on the matter :) Feb 05 15:03:04 JPEW: specifically which piece of archiver? Feb 05 15:03:28 rburton: I wondered when you'd read that :) Feb 05 15:04:28 RP: INHERIT += "archiver" ARCHIVER_MODE[dumpdata] = "1" ARCHIVER_MODE[recipe] = "1" Feb 05 15:04:58 RP: We use it to generate our archive of open source software that we post for compliance Feb 05 15:09:30 JPEW: thanks, its useful to know which options are actually used Feb 05 15:17:38 If anyone has anything they want to highlight in this week's status, let me know Feb 05 15:32:39 Has anyone been having strange busybox errors? Feb 05 15:33:13 build issue with ptest enabled Feb 05 15:34:50 This would be a runtime error. Turns out, busybox 1.29.2 had an error that prevented it from correctly parsing config file. https://git.busybox.net/busybox/commit/?h=1_29_3&id=2993551efcb09ace35b40db608ab08159834692b Feb 05 15:35:23 heh Feb 05 15:35:53 They fixed in in 1.29.3 (which is the only change from 1.29.2 -> 1.29.3). I sent a patch to upgrade the recipe to the mailing list (also, affects thud) Feb 05 15:36:57 JPEW: can you post the patch for thud please Feb 05 15:36:59 pretty please Feb 05 15:37:05 so ptest did its job? Feb 05 15:37:11 * armpit thud, the sound my head makes hitting the desk Feb 05 15:38:08 Is http://lists.openembedded.org/pipermail/openembedded-core/2019-February/278727.html sufficent, or do you want a separate e-mail for thud? Feb 05 15:38:11 rburton, if the update is a bug fix only, its fine Feb 05 15:45:14 RP: lack of interest in package updates :( Feb 05 15:46:51 kanavin: I have that in there Feb 05 15:47:15 kanavin: I think we need to send an email to the maintainers and ask if there is a problem Feb 05 15:47:39 RP: yes, something like that Feb 05 15:48:35 kanavin: rburton technically owns this but basically has other problems so I guess this would fall to me Feb 05 15:48:48 kanavin: We've given up on mut atm too :( Feb 05 15:49:20 RP: I suppose Ross is busy with not-yocto things Feb 05 15:50:30 indeed. good news is the thing has a deadline. Feb 05 15:56:15 rburton: I'm hoping we get you back! ;-) Feb 05 15:59:37 YPTM: Bridge is with Zoom at: https://zoom.us/j/990892712 Feb 05 15:59:47 YPTM: Sjolley joined Feb 05 16:00:12 YPTM: Joshua Watt here Feb 05 16:00:26 YPTM -- Randy joined Feb 05 16:00:50 kergoth, rburton: just wanted to say thanks for your patience in helping me out yesterday (re: version/revision numbering) Feb 05 16:01:22 np Feb 05 16:01:37 versioning is trivial if you've done distro work before but if not, then it can be a mystery Feb 05 16:01:38 YPTM: David Reyna joined Feb 05 16:02:04 hmm, bridge isn't liking me Feb 05 16:03:31 getting a "System Error" Feb 05 16:04:19 it's sitting and spinning on 'Connecting...' here Feb 05 16:04:50 trying to join from the Browser I get "502 Bad Gateway" Feb 05 16:05:38 I'm getting the same, 502 Bad Gateway when attempting to open the zoom conference call url Feb 05 16:06:12 sjolley: think something has failed in the system :( Feb 05 16:07:16 No problem. I'll follow along by watching the Google Doc :) Feb 05 16:07:42 it just let me in Feb 05 16:07:46 ...maybe... Feb 05 16:08:07 it loaded, but I'm by myself.. Feb 05 16:08:11 I think zoom itself is having problems Feb 05 16:08:40 ok, it let me in now.. Feb 05 16:08:47 oh cool. I just got into the voice bridge too. Feb 05 16:08:50 RP, you on ? Feb 05 16:09:11 YPTM: good morning, Nick is on the call Feb 05 16:09:11 Is it you in yet? Feb 05 16:09:14 RP? Feb 05 16:09:31 * RP can't get audio working Feb 05 16:10:10 We can hear you. Feb 05 16:20:04 RP: you dropped. Feb 05 16:20:06 can anyone hear me ? Feb 05 16:20:08 did we lose audio Feb 05 16:20:11 armpit: no Feb 05 16:20:13 yes Feb 05 16:20:19 vmeson, k Feb 05 16:20:31 I got kicked out Feb 05 16:20:33 Yocto cone of silence Feb 05 16:20:37 won't let me in :( Feb 05 16:20:45 I got kicked also Feb 05 16:21:11 Not sure what is happening in Zoom Feb 05 16:21:35 sjolley: can you rejoin? Feb 05 16:21:52 I rejoined Feb 05 16:21:53 i was able to rejoin Feb 05 16:22:45 * RP is back on Feb 05 16:23:05 I rejoined Feb 05 16:35:23 YPTM: Is over Feb 05 16:45:33 New news from stackoverflow: Yocto + boost. Is there any way to get libboost.a into the SDK? Feb 05 17:18:30 https://juliareda.eu/2019/02/article-13-worse/ is mind boggling. EU folk, email your MEPs Feb 05 17:23:55 kergoth: not sure we'll have MEPs soon... Feb 05 17:26:09 heh, true Feb 05 17:26:16 * kergoth shakes head Feb 05 17:30:57 Heya, Ive got a standalone kernel module that compiles and works fine (in its own layer). However when I modify the kernel it builds against, it seems theres no dependency to force it to rebuild. Is there a way to create a dependency against "virtual/kernel"? Feb 05 17:32:35 The .bb file does include 'DEPENDS += "virtual/kernel"' but this doesn't seem to be enough to force a recompile... Feb 05 18:15:03 derRichard: can you pm me please? Feb 05 18:19:21 no_such_user: you need source dep may be Feb 05 18:31:00 Hi Feb 05 18:32:04 If II have followiind setup buxybox.bb <-- busybox.bbappend (bsp layer <-- busybox.bbappend (my layer)) Feb 05 18:32:47 in both bbappend files I have do_install_append Feb 05 18:33:03 but for some reason file from my layer is not used Feb 05 18:34:49 It looks like my busybox.bbappend is executed before this from bsp layer Feb 05 18:35:21 how to change this order that my recibe will be at the end Feb 05 18:48:46 derRichard: you there? Feb 05 18:55:58 why am i getting this after a "devtool reset -a"? ... Feb 05 18:56:09 ERROR: recipe linux-variscite is already in your workspace Feb 05 19:00:51 nm Feb 05 19:01:21 sumpin' got hosed, logging out and back in, resourcing oe-init-build-env fixed Feb 05 19:04:16 no, it did not... Feb 05 19:06:46 is anyone building for x86 hardware targets? (not just qemu) Feb 05 19:15:10 nerdboy: yeah Feb 05 19:15:38 mind sharing witch layers/branches? Feb 05 19:15:49 mind sharing which layers/branches? Feb 05 19:16:07 brain typos... Feb 05 19:16:15 well.. i mean its only for my own personal laptop Feb 05 19:16:33 yeah, i don't quite mean that specific... Feb 05 19:16:47 :-D Feb 05 19:17:05 last i tried, the images would boot but were missing things like network/display/other Feb 05 19:17:31 I use connman Feb 05 19:17:35 *booting on vendor hardware like NCR PoS terminals Feb 05 19:19:53 I build for real hardware all the time.. but the combination of layers changes a lot from customer to customer.. Feb 05 19:20:03 there is no single combination... Feb 05 19:20:10 can you share your bblayers/local configs? i made a foss-intel-bsp manifest thing, tested musl/selinux Feb 05 19:20:44 My configurations are based on Wind River Linux... we do NOT use intel BSPs.. we use kernel.org/YP kernel sources for our BSPs.. Feb 05 19:20:48 kroon: ^^ Feb 05 19:21:14 why is devtool reset -a not resetting all recipes? Feb 05 19:21:37 i do have meta-intel and meta-intel-contrib in there, should i avoid those? Feb 05 19:21:49 I don't use them, I need a standard kernel not a vendor specific kernel.. Feb 05 19:22:03 i normally use my own kernel recipes Feb 05 19:22:04 doesn't matter if it's Intel, NXP, or anyone else.. I don't want vendor specific kernels Feb 05 19:22:23 meaning not vendor kernels... Feb 05 19:22:24 nerdboy: I just use standard OE-Core, meta-oe/meta-perl + custom layer for doing personal specializations Feb 05 19:22:45 A full configuration for me has about 44 layers enabled.. (note, I'm NOT using software from all of them) Feb 05 19:23:50 oe-core, meta-openembedded, meta-security, meta-selinux, meta-virtualization, meta-realtime, meta-yocto, meta-gplv2, meta-secure-core, meta-cloud-services, meta-intel, meta-freesale, meta cgl, meta-dpdk, meta-intel-qat ... (and most of those contain sublayers as well) Feb 05 19:23:51 nerdboy: i suspect I should also have meta-intel to get CPU ucode updates (which I assume are in there) Feb 05 19:24:35 keep in mind that microcode applied in USERSPACE means that many of th security fixes and other hardware fault fixes are not actually applied.. Feb 05 19:24:50 to be fully 'fixed', you need your bios vendor to apply the latest microcode fixes for your CPU Feb 05 19:25:00 the userspace microcode update will fix 'some' things.. Feb 05 19:25:02 so either oe-core/my kernel or poky/yp kernels? Feb 05 19:25:14 it's possible for the kernel to apply a microcode early as well.. that is 'better' then userspace, but still not complete Feb 05 19:25:31 oe-core and yp are the same when we talk about kernels.. Feb 05 19:25:46 so it's oe vs 'custom'.. custom can be provided by a semi's layer, you, etc.. Feb 05 19:26:07 for -my- purposes, I need a common kernel and want the OE / YP kernel as the source for my work Feb 05 19:26:24 but my purposes may not be right for everyone, so sometimes choosing the semi-vendor kernel may make more sense to you Feb 05 19:26:30 (or even using a raw kernel.org kernel) Feb 05 19:27:04 no vendor kernels in this case Feb 05 19:27:57 fray: do you have an example somewhere i can poke at? or is it all customer-specific stuff? Feb 05 19:28:36 and what bits of meta-intel do you actually use? Feb 05 19:31:54 you on a particular version? thud? master? Feb 05 19:34:07 i wasn't sure what to try, this is the last layer config: https://bpaste.net/show/6ee3a432890c Feb 05 19:34:29 i'm *very* open to suggestions Feb 05 19:34:43 what version of OE/YP are you using, which branch? Feb 05 19:35:28 I've got public 2.4, 2.6 and master stuff... I just need to know which to point you att Feb 05 19:35:46 looks like the last one i tried was sumo Feb 05 19:36:03 how about 2.6? Feb 05 19:36:17 Sumo looks like 2.5.. so 2.6 is the clsoest I have.. Thud Feb 05 19:36:32 https://github.com/WindRiver-Labs/wrlinux-x Feb 05 19:36:35 thats where to start.. Feb 05 19:36:44 clone that down, use the branch WRLINUX_10_18_BASE Feb 05 19:37:04 be sure to read the README on github BEFORE cloning.. Feb 05 19:37:12 since there is a specific sequence you need to follow.. Feb 05 19:37:44 in the READE, use WRLINUX_10_18_BASE as the branch, and https://github.com/WindRiver-Labs/wrlinux-x as the 'repoURL' (you can change https to http or git and it'll work fine) Feb 05 19:38:26 then run the setup tool, ./wrlinux-x/setup.sh --all-layers and you'll get the configuration I start with when I don't know specifics about the project yet Feb 05 19:41:09 Does the "psmic" recipe actually build on master ? Feb 05 19:41:23 I don't know if I've tried it lately.. Feb 05 19:41:55 I get an error about missing autopoint Feb 05 19:42:45 thanks Feb 05 19:42:49 it build if I remove the EXTRA_AUTORECONF_remove Feb 05 19:43:25 s/psmic/psmisc Feb 05 20:42:22 * nerdboy goes back to ansible hacking for a bit... Feb 05 21:29:18 armpit: your last reply to yocto@lists.yoctoproject.org didn't go through since it came with your @gmail address. do you want to resend or should i let it go to the list? Feb 05 21:30:23 ndec, can you let it go please Feb 05 21:31:51 done. Feb 05 21:32:32 thanks Feb 05 21:33:47 gah, had the feeling something was wrong in the server reconnect code, think I've found it Feb 05 21:33:55 armpit: suspect stable will need this bitbake fix Feb 05 21:34:27 RP, who you calling stable ; ) Feb 05 21:38:47 armpit: are any of us? :) Feb 05 21:44:51 Hi, I've got a quick question: I would like to set a global variable in distro.conf to the value of the current git revision of my distro layer. Is it possible to define and run a python or shell function inside a .conf file and use that to get the git revision? Feb 05 21:51:17 robbawebba: its better to feed it from your env setup script Feb 05 21:56:48 you can look into IMG_VERSION setting in Yoe distro https://github.com/YoeDistro/yoe-distro/commit/d8f8e0efb5bc09f65a6fe86af9fca21872caf703 Feb 05 22:08:41 khem: ahh nice, thank you for that example. It seems that yoe distro has a pretty big script system, whereas I usually just rely on the poky/oe-init-build-env script for environment setup. Feb 05 22:08:59 Do you know if it's possible to have my own script run as part of oe-init-build-env? Feb 05 22:19:11 robbawebba: no but you can write your own script. note that oe-init-build-env just calls another script, so your script can do the same Feb 05 22:20:42 derRichard: still not around? (vacation day?) Feb 05 22:21:41 robbawebba: yes, it tries to address lot of small quirks that people often run into. Feb 05 23:10:39 kergoth: i was going to share something ugly i did with you but i can't /query you Feb 05 23:30:42 kanavin: https://autobuilder.yoctoproject.org/typhoon/#/builders/85/builds/198 had python in it, still a few issues Feb 05 23:37:48 kanavin: qemuarm isn't yours, multilib 5b is, need to look at selftest in detail **** ENDING LOGGING AT Wed Feb 06 02:59:57 2019