**** BEGIN LOGGING AT Thu Jul 18 02:59:56 2019 Jul 18 07:22:19 Hi, I'm trying to add portaudio-v19 recipe in my yocto. http://cgit.openembedded.org/meta-openembedded/tree/meta-oe/recipes-support/portaudio/portaudio-v19_20161030.bb?h=master. do_compile failing. meta/recipes-support/portaudio/portaudio-v19_20161030.bb:do_compile) failed with exit code '1' . | arm-oe-linux-gnueabi-libtool: compile: unable to infer tagged configuration Jul 18 07:22:19 | arm-oe-linux-gnueabi-libtool: error: specify a tag with '--tag' Jul 18 07:22:19 | make: *** [src/common/pa_cpuload.lo] Error 1 Jul 18 07:22:19 | make: *** Waiting for unfinished jobs.... Jul 18 07:22:19 | arm-oe-linux-gnueabi-libtool: error: specify a tag with '--tag' Jul 18 07:22:21 | make: *** [src/common/pa_process.lo] Error 1 Jul 18 07:22:23 | arm-oe-linux-gnueabi-libtool: error: specify a tag with '--tag' Jul 18 07:22:25 | make: *** [src/os/unix/pa_unix_util.lo] Error 1 Jul 18 07:22:27 . Jul 18 07:52:08 "Your membership in the mailing list yocto has been disabled due to Jul 18 07:52:09 excessive bounces" Jul 18 07:52:44 is anyone else seeing this, maybe a problem on yocto side? Jul 18 08:22:59 mcfrisk: its possible, we've been having a lot of mailing list problems :( halstead could probably comment Jul 18 09:15:31 Does anybody use headless EGL over GBM here? I have an opengl app using egl context, it works fine under Arch, but doesn't work with Sato without X11 or wayland (egl not initialized error) Jul 18 09:15:58 Which features do I need for headless egl? Jul 18 09:16:55 There is card's node in /dev/dri Jul 18 09:31:19 hello folks Jul 18 09:31:36 i need the meta-printing layer from the layer index because it contains the python-reportlab recipe Jul 18 09:31:43 but it is not provided for branch warrior, apparently Jul 18 09:32:04 bitbake-layers layerindex-fetch -n -b warrior meta-printing says there's no such layer. but if specifying -b rocko, it works Jul 18 09:32:08 what's recommended in this case? Jul 18 09:33:09 is there master branch with newer than rocko changes? Jul 18 09:33:27 it would be closer to warrior than old rocko Jul 18 09:34:15 it also has sumo Jul 18 09:34:23 ah, it has master. I'll use that then Jul 18 09:34:27 thanks Jul 18 09:36:58 JaMa, it seems to have moved to "meta/" as a core-layer Jul 18 09:38:57 ah, false. it just printed meta/ as a dependency haha Jul 18 09:44:17 JaMa, its master README says it's either compatible with master or krogoth. but the repo also has sumo and rocko branches. my understanding is that both of those are more recent than krogoth. Jul 18 09:44:23 sounds to me like the README is outdated Jul 18 09:45:54 New news from stackoverflow: How to create a initramfs image with uImage and cpio rootfs? Jul 18 09:47:56 litb: yes README is probably outdated, just check git log to see which branch has most recent changes Jul 18 10:17:23 oh is that my meta-printing Jul 18 10:17:25 litb: ^ Jul 18 10:17:38 its dead and everything in there should be somewhere else Jul 18 10:18:19 if there isn't then just copy the recipe out and contribute it somewhere better Jul 18 10:18:56 or, take ownership of the layer and update it. it was a start at consolidating printing pieces from across the layers that didn't take off Jul 18 10:19:05 like it has a really old cups in Jul 18 10:22:42 ah i see. Jul 18 10:24:55 thanks. while working with setuptools, I found that it ignores the TUNE_ARCH settings. my tune is corei7-32 , buht setuptools compiles python libs to 64bit. i.e it generates lib.linux-x86_64-2.7/ Jul 18 10:25:12 rburton_, looks like these python setup scripts don't work well with cross compilation Jul 18 10:25:56 so, things fail once it tries to find libjpeg.so because it ignores the 32bit libraries that yocto built for it as part of DEPENDS resolving Jul 18 10:42:48 yeah a lot of C/python setup is an absolute horror Jul 18 10:42:55 to be fair, python itself is a horror to cross compile Jul 18 11:21:31 rburton_, maybe I can ignore that the lib folder's name is incorrect. after all it's just a folder name. it generates those by using uname Jul 18 11:22:28 rburton_, but, why does "your" python-pillow look for libs in Looking in z for ['/mnt/vcs/git/poky/build/tmp/work/corei7-32-FOO-linux/python-pillow/3.3.0-r0/recipe-sysroot-native/usr/lib', '/mnt/vcs/git/poky/build/tmp/work/corei7-32-FOO-linux/python-pillow/3.3.0-r0/recipe-sysroot/lib'] ? Jul 18 11:23:09 it's not supposed to look into recipe-sysroot-native when it builds the normal variant of the package. or am I missing something? moreover it is incosistent here, because it looks into both kind of sysroots. hmm Jul 18 11:23:38 that's why it fails to compile: because the libs are in recipe-sysroot/usr/lib not in recipe-sysroot/lib. but it only looks in lib/ Jul 18 11:26:57 rburton_, maybe it has to do with the sed lines that you do in do_compile_prepend? for example it does -e s:/usr/lib:${STAGING_LIBDIR}:g . However I can't see how that would bring recipe-sysroot-native into play and how it removes /usr/lib from consideration at all Jul 18 11:34:09 Is there a comprehensive documentation on bitbake recipes? Like what variables generally exist and what they mean, what operators do exactly what? Jul 18 11:34:09 Maybe even includes some popular bbclasses? Jul 18 11:34:09 It's possible, but rather tedious to learn by reading random third-party recipes ... Jul 18 11:37:31 jmiehe: maybe https://www.yoctoproject.org/docs/latest/ref-manual/ref-manual.html ? Jul 18 11:52:10 jmiehe, I found that the bitbake manual is easier to understand than the yocto manual Jul 18 11:52:23 because it's more dense on bitbake concepts without the yocto concepts ontop Jul 18 12:09:29 In https://www.yoctoproject.org/docs/latest/bitbake-user-manual/bitbake-user-manual.html#override-style-operation-advantages -- the "+=" attempt fails because "inherit" is before setting FOO? did I get that right? Jul 18 13:33:16 When I add python3 to DEPENDS I can see the libraries and modules added to recipe-sysroot, but not the actual python3 binary Jul 18 13:33:57 $ find ../recipe-sysroot/ -iname python3 -type f../recipe-sysroot/sysroot-providers/python3 Jul 18 13:34:22 Sorry, those lines are: Jul 18 13:34:31 $ find ../recipe-sysroot/ -iname python3 -type f Jul 18 13:34:35 ../recipe-sysroot/sysroot-providers/python3 Jul 18 13:35:21 And I need the python3 binary in there because the configure step uses which to find the python3 binary Jul 18 13:35:38 Any ideas what might I be doing wrong? Jul 18 13:36:08 I think you need `inherit python3native` Jul 18 13:37:22 DEPENDS = "python3" would be if you needed to compile/link against the target python (as one example) Jul 18 13:44:38 JPEW: So python3-native if I need the actual binary and python3 if I need the target's python3 libraries Jul 18 13:46:56 yeah Jul 18 13:47:21 no point putting target binaries in the sysroot as you can't run them, so the target sysroot doesn't have a /usr/bin at all Jul 18 13:47:29 its essentially just headers and libraries Jul 18 13:47:52 of course if you want to run py3 there's a perfectly good one on $PATH as you're running bitbake already, which i py3 Jul 18 13:48:25 if configure is trying to find py3 and its cmake then you can beat it with a stick to shut it up Jul 18 13:48:53 EXTRA_OECMAKE += "-DPYTHON_EXECUTABLE=${HOSTTOOLS_DIR}/python3" Jul 18 13:49:09 if its not cmake then it will usually just find the one on $PATH for you Jul 18 14:11:27 rburton_: Thanks, I understand now why the binary is not there. It is not cmake but using ${HOSTTOOLS_DIR}/python3 would be helpful here Jul 18 15:13:34 JPEW: we have another bzip meta-mingw version mismatch in -next :( Jul 18 15:19:29 The latest version of bzip should include the MinGW patch and the .bbappend in meta-mingw can be deleted Jul 18 15:27:18 RP: I removed bzip2 from master-next of meta-mingw. That should allow you to test it out. Jul 18 15:27:35 Or rather keep it from failing on the autobuilder Jul 18 15:28:33 JPEW: thanks! Nice to be able to drop more patches Jul 18 15:30:19 JPEW: test build fired which includes that :) Jul 18 15:30:24 Yep. Upstream bzip2 has been pretty responsive.... We might be able to help them use a standardized test format as part of https://bugzilla.yoctoproject.org/show_bug.cgi?id=13444 Jul 18 15:30:25 Bug 13444: normal, Medium+, 2.99, newcomer, NEW , Add test suite for bzip2 Jul 18 15:30:47 JPEW: that would be very cool Jul 18 17:42:41 Intel folks: I have a device on which Im trying to boot poky on, and by device I mean my laptop Jul 18 17:42:54 Im building master on meta-intel as well Jul 18 17:43:20 after the bootloader, when i select a boot option the screen goes blank and I never see anything after that Jul 18 17:44:00 I'm quite certain that it boots, since even the keyboard reponds, basically if I blindly login and do something like poweroff or reboot it reponds Jul 18 17:44:34 I seem to remember we had an issue with the Minnowboard on which we could only see boot messages on either the serial terminal or HDMI Jul 18 17:44:53 sadly the laptop does not have a serial terminal so I cant say Jul 18 17:45:17 I've tried meddling around with several ttys with no success Jul 18 17:45:46 and i also suspected it could be because of the i915 driver, but as far as I can see it it enabled through meta-intel Jul 18 17:45:48 any clues? Jul 18 17:56:20 aehs29: how odd. try blacklisting the i915 kernel driver? Jul 18 17:56:49 aehs29: is it a shiny new laptop with really new chipset? Jul 18 18:00:08 rburton_: no its from a couple of years ago Jul 18 18:00:17 there goes that theory then Jul 18 18:00:24 rburton_: well the thing is that other distros boot fine, and I can see that i915 is being loaded Jul 18 18:00:28 on those distros Jul 18 18:00:54 maybe try warrior just to see if its a recent change? Jul 18 18:01:16 deffo worth blacklisting the driver from grub to see if that is the problem Jul 18 18:01:34 rburton_: I'm not sure what magic happens on live booting on those distros that is perhaps able to detect what it needs and adapt Jul 18 18:01:43 rburton_: i915.modeset=0? Jul 18 18:02:25 nuclear option: module_blacklist=i915 in the grub menu Jul 18 18:03:02 rburton_: wuuuu Jul 18 18:03:08 rburton_: new information Jul 18 18:03:16 rburton_: core-image-sato booted just fine Jul 18 18:03:33 ah. was it a core-image-minimal or something before? Jul 18 18:03:37 yup Jul 18 18:04:13 i bet we did something like turn off the traditional console in the kernel config but no i915 drivers mean nothing works :) Jul 18 18:04:14 bios not efi? Jul 18 18:04:32 i can't remember if i turned the efi console on in meta-acrn, or meta-intel... Jul 18 18:04:40 bios Jul 18 18:05:41 can you file a bug please. there's probably been a bit too much modularising or something is off when it shouldn't be. Jul 18 18:06:17 I can, I mean its not like we support this laptop but yes Jul 18 18:06:35 also now that it booted on sato I can debug a little further Jul 18 20:03:34 aehs29: meta-intel should work on most typical hardware though Jul 18 21:26:35 rburton_: define typical haha Jul 18 22:13:48 aehs29: there lies a topic of years of debate :) Jul 18 22:17:55 New news from stackoverflow: OpenBMC with Raspberry Pi (2 or 3) and build bmcweb? Jul 18 23:40:17 Good evening. Jul 18 23:55:14 I'm messing with this intel aero drone. It runs a yocto instlll. Optionally ubuntu, but after I install ubuntu the keyboard and mouse don't work. The problem I'm having is that I can't keep a laptop connected to the aero's wifi. The aero runs as its own AP, but after about 1400 pings the laptop loses connection. I'm looking at the logs on the aero to see what I can see, anyone know what the process name is for the wifi in yocto? Jul 19 00:38:21 puff: no idea what that image uses to connect, probably connman?, it could also be using the network stack itself from iw Jul 19 00:48:17 New news from stackoverflow: Yocto "Failed to run qemu: Could not initialize SDL(x11 not > available)" Jul 19 01:20:04 http://ix.io/1OS2 Jul 19 01:27:20 puff: yeah that looks like its coming from the driver itself, not the userspace utility Jul 19 01:27:39 puff: next question is what kernel and driver versions are you using and are you able to upgrade those Jul 19 01:46:24 Linux intel-aero 4.4.113-yocto-standard #1 SMP PREEMPT Tue May 22 11:10:22 IST 2018 x86_64 x86_64 x86_64 GNU/Linux Jul 19 01:46:45 I'm game for upgrading it, I've reinstalled yocto on this thing about a dozen times now so worst case I just reinstall yocto. Jul 19 01:49:00 http://ix.io/1OS8 Jul 19 01:49:00 Jul 19 02:10:32 puff: since it says yocto-standard then it means you are using linux-yocto Jul 19 02:11:26 puff: I would suggest you upgrade, I'm not saying it'll be easy since we used 4.4 like 3 years ago Jul 19 02:12:32 puff: start by checking what yocto release you are using, it'll depend on that if its worth updating the whole release or just the kernel Jul 19 02:13:23 puff: for example, rocko supports 4.4, 4.9, 4.10 and 4.12, so maybe its just a matter of changing the PREFERRED_VERSION (hopefully) Jul 19 02:13:53 puff: warrior(the current release) supports 4.19 and 5.0 Jul 19 02:49:59 aehs29: Ah, it's standard? Interesting. I thought it was some special yocto for these drones. Jul 19 02:52:19 https://downloadcenter.intel.com/download/27833/Intel-Aero-Platform-for-UAVs-Installation-Files?v=t Jul 19 02:56:26 Hm, how do I check the release? Google is failing me here. **** ENDING LOGGING AT Fri Jul 19 02:59:57 2019