**** BEGIN LOGGING AT Thu Jan 11 03:00:02 2018
Jan 11 05:13:29 <droman0> I need to include a build timestamp for traceability and so I included the os-release package in my image recipe however I noticed that if I cat /etc/os-release, the BUILD_ID variable shows me GMT time. How can I change to wall time?
Jan 11 09:05:26 <yocti> New news from stackoverflow: Yocto kernel module path <https://stackoverflow.com/questions/48203001/yocto-kernel-module-path>
Jan 11 09:57:47 <eduardas_m> hello, I get an openssl: not found error even though I have a DEPENDS += " openssl-native" in my recipe. What am I doing wrong?
Jan 11 09:59:11 <eduardas_m> this is my recipe: https://pastebin.com/E65rcFz4
Jan 11 10:00:51 <eduardas_m> I am attempting to put an update.sh script in my deploy directory and the encrypt it for further use in an SWU update archive (meta-swupdate)
Jan 11 10:05:41 <yocti> New news from stackoverflow: downgrading QT-5.7 to QT-5.4 in Yocto( morty 2.2) for nitrogen6sx <https://stackoverflow.com/questions/48061439/downgrading-qt-5-7-to-qt-5-4-in-yocto-morty-2-2-for-nitrogen6sx>
Jan 11 10:35:47 <yocti> New news from stackoverflow: lcd interface to linux 4.12.4 linux kernel for imx6ul <https://stackoverflow.com/questions/48204482/lcd-interface-to-linux-4-12-4-linux-kernel-for-imx6ul>
Jan 11 11:43:27 <eduardas_m> hello, I just noticed that IRC logs for this channel are not available for days since the 1st of January this year... last entry is 2017-12-31
Jan 11 11:44:06 <rburton> halstead: ^ irc logs for 2018 are missing
Jan 11 12:02:52 <eduardas_m> rburton: I get an openssl: not found error for the encrypt_script task of my recipe on the first bitbake run: https://pastebin.com/ugUJm8ZF
Jan 11 12:03:09 <eduardas_m> recipe completes on second bitbake run though
Jan 11 12:03:17 <eduardas_m> this is with Yocto Rocko
Jan 11 12:03:41 <eduardas_m> this is weird since I thought depending on openssl-native is enough
Jan 11 13:31:42 <acrap> hi, folks! Does MACHINE_OVERRIDE equal MACHINE name from local.conf?
Jan 11 13:39:35 <LetoThe2nd> acrap: seen somewhere? AFAIK this is not a generic variable
Jan 11 13:40:13 <LetoThe2nd> or are you talking about http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#var-MACHINEOVERRIDES
Jan 11 13:56:16 <tasslehoff> my recipe needs a CROSS_COMPILE variable set, with path and prefix to compiler. what is the proper way to set it?
Jan 11 14:11:25 <rburton> eduardas_m: they're racing with do_populate_recipe_sysroot
Jan 11 14:11:36 <rburton> eduardas_m: the sysroot is only populated once that has ran
Jan 11 14:11:55 <rburton> eduardas_m: make it "after do_prepare_recipe_sysroot'
Jan 11 14:14:32 <Adam___> Has anyone used "dfu tftp" in u-boot? If so, is there any size limitation on the rootfs image?
Jan 11 14:36:29 <eduardas_m> rburton: thank you a lot... that actually helped
Jan 11 14:37:08 <eduardas_m> rburton: but that makes me wonder: how did you know that? how does one figure out that there is a race condition?
Jan 11 14:37:48 <kanavin> eduardas_m: an issue that disappears on the second run is almost certainly a race
Jan 11 14:38:06 <rburton> eduardas_m: 1) because you described a race (doesnt work then does work) 2) because your task is only saying "before build" so won't expect a sysroot to exist
Jan 11 14:38:28 <rburton> (plenty of tasks are valid without a sysroot, but you're trying to run a command from it, so yeah you need a sysroot)
Jan 11 14:38:39 <eduardas_m> rburton, kanavin: is there a typical way for one to debug race conditions with bitbake?
Jan 11 14:40:23 <rburton> eduardas_m: the only typical race is sysroot contents, which thanks to recipe-specific-sysroots in rocko is "exists" or "doesn't exist", so thats trivial
Jan 11 14:40:48 <rburton> not like the old days where configure would say a binary is present and compile wouldn't find it, or even more fun when you're trying to run a binary that another task is writing to
Jan 11 14:41:36 <kanavin> eduardas_m: much more typical races we encounter are in the upstream source code - poorly written Makefiles etc.
Jan 11 14:46:23 <eduardas_m> rburton, kanavin: that was really informative, thank you
Jan 11 14:46:39 <luneff> hey guys! what does one do on /bitbake/lib/bb/runqueue.py:1179: ResourceWarning: unclosed file <_io.BufferedWriter name=29> build crash?
Jan 11 14:47:00 <luneff> I'm building meta-toolchain-qt5, rocko version
Jan 11 14:49:06 <kanavin> luneff: does it actually crash?
Jan 11 14:49:17 <luneff> it does exit all building threads, so yes
Jan 11 14:49:34 <luneff> rerunning everything with BB_NUMBER_THREADS=1
Jan 11 14:49:47 <luneff> to check if that solves the problem
Jan 11 14:49:52 <kanavin> luneff: you might need to increase kernel limits, but I can't give more specific info
Jan 11 14:51:14 <luneff> oh, indeed it could be ulimit problem, thanks! 1024 could be not enough for 16-thread build
Jan 11 14:52:02 <kanavin> luneff: ak@kanavin-desktop:~/development/poky/build/workspace/sources/webkitgtk$ ulimit -n
Jan 11 14:52:02 <kanavin> 1024
Jan 11 14:52:19 <kanavin> and I haven't seen this
Jan 11 14:52:28 <kanavin> with 42 threads ;)
Jan 11 14:55:04 <luneff> 2048 didn't help be as well. It looks more like a recipe problem then
Jan 11 14:57:41 <luneff> pity that, as my life is slightly worse without the SDK :-(
Jan 11 15:00:12 <luneff> will continue to build in a single thread until it is clear what causes that fail
Jan 11 15:12:14 <rburton> luneff: resourcewarning is just a warning, that's not the crash
Jan 11 15:16:44 <luneff> rburton, after this warning all threads have quit, citing "unexpected worker exit". I've passed through some stage with BB_NUMBER_THREADS=1 and it seems to build properly after that
Jan 11 15:18:12 <luneff> well, not really, as qtlocation build fails :-D
Jan 11 15:19:42 <kanavin> luneff: how much RAM do you have?
Jan 11 15:19:56 <luneff> 16 gigs
Jan 11 15:20:01 <kanavin> luneff: quadruple that :)
Jan 11 15:20:06 <luneff> O_O
Jan 11 15:20:28 <kanavin> or stop using C++ frameworks
Jan 11 15:20:38 <kanavin> they eat RAM like nothing else at compile time
Jan 11 15:21:37 <luneff> qtlocation build fail: https://pastebin.com/ihX9Hdam
Jan 11 15:21:50 <luneff> actually, I think it is more like Qt isn't prepared to meet gcc 7.1
Jan 11 15:22:24 <kanavin> "arm-poky-linux-gnueabi-g++: internal compiler error: Killed (program cc1plus)"
Jan 11 15:22:29 <kanavin> that's out of ram :)
Jan 11 15:23:06 <luneff> oh gods
Jan 11 15:23:54 <kanavin> luneff: instead of BB_NUMBER_THREADS, use a setting to limit parallel make execution, that should help
Jan 11 15:23:59 <kanavin> I forgot what it's called...
Jan 11 15:24:07 <kanavin> obviously that will make the build slower
Jan 11 15:24:16 <luneff> PARALLEL_MAKE
Jan 11 15:25:07 <kanavin> you don't generally want to disable parallel make altogether though, as you will spend your life waiting for builds to complete
Jan 11 15:25:12 <kanavin> just reduce it until it works
Jan 11 15:25:21 <luneff> yes, thanks! I'll try
Jan 11 15:25:39 <kanavin> and get management to buy you a shiny workstation :)
Jan 11 15:25:54 <luneff> I remember days when I compiled centericq on a FreeBSD 4.11 on a P166MMX with 16 MBs of RAM. It was C++. It took 27 hours to compile an ICQ program :-D
Jan 11 15:26:16 <luneff> could be the same story here :-D but 15 years later
Jan 11 15:31:08 <kanavin> luneff: I wonder how usable Go's Qt bindings are...
Jan 11 15:31:49 <luneff> kanavin, dunno. saw ruby bindings as well. AFAIK, only PyQt is 100% complete and safe
Jan 11 15:32:32 <luneff> although, with QML, the majority of the program could be written in Javascript, regardless :-)
Jan 11 15:32:41 <kanavin> luneff: I wouldn't mind if Go squeezes out C and C++ for most use cases
Jan 11 15:32:54 <luneff> me too :-)
Jan 11 16:37:15 <yocti> New news from stackoverflow: Yocto doesn't pack busybox syslog files <https://stackoverflow.com/questions/48211589/yocto-doesnt-pack-busybox-syslog-files>
Jan 11 16:51:02 <xthunderheartx> luneff: "I remember days when I compiled centericq on a FreeBSD 4.11 on a P166MMX with 16 MBs of RAM."
Jan 11 16:52:14 <xthunderheartx> Pahhhh! I remember installing Linux on a 386SX with an amber monitor.
Jan 11 16:53:27 <xthunderheartx> And yes, I have loaded systems from mag *and* paper tape.
Jan 11 16:56:31 <xthunderheartx> Had to walk through 9 miles of snow to do it too ... whipper snappers ...
Jan 11 17:37:12 <batman_> hi, Are there functions like getVar/setVar to access datastore variables frim within a bash function?
Jan 11 17:38:12 <batman_> Right now, if I set a variable defined in a bbclass file from within a bash function and immediately print it, the new value is not getting reflected.
Jan 11 17:38:49 <kergoth> inline python with ${@} is processed at variable expansion time, not run time
Jan 11 17:42:39 <batman_> @kergoth, right now I am setting variables with something like this PROC=$(grep -r "Processor.*${cpu_id} .....)
Jan 11 17:43:16 <batman_> Should I be instead using ${@setVar(PROC, ....} ?
Jan 11 17:43:27 <batman_> Or is setVar only for python functions?
Jan 11 17:49:30 <halstead> rburton, eduardas_m 2018 logs are in the web interface now. Thanks.
Jan 11 18:37:37 <yocti> New news from stackoverflow: How do I update a datastore variable from inside a bash variable? <https://stackoverflow.com/questions/48213497/how-do-i-update-a-datastore-variable-from-inside-a-bash-variable>
Jan 11 19:50:03 <denix> is there a way in RSS to stage other package's sysroot w/o adding it to DEPENDS?
Jan 11 20:40:49 <smurray> denix: iirc there’s a hook for tools for packaging, forget the var name. That might not do what you need, I suspect
Jan 11 20:59:40 <denix> smurray: thanks. let me know if you remember the name
Jan 11 21:24:44 <smurray> denix: http://www.yoctoproject.org/docs/2.4.1/ref-manual/ref-manual.html#var-PACKAGE_WRITE_DEPS
Jan 11 21:25:55 <smurray> denix: it's supposed to be used for package pre and post install deps.  In AGL we needed it for something related to SMACK iirc
Jan 11 21:28:57 <denix> smurray: thanks! don't think it will work for my need
Jan 11 21:29:10 <smurray> denix: yeah, I figured
Jan 11 21:50:20 <wasutton3_work> so im trying to build yocto rocko with a pine64 layer, however u-boot fails to build citing "/bin/sh:1: arm-linux-gnueabihf-gcc: not found", however arm-linux-gnueabihf-gcc is in my local path and is executable
Jan 11 21:54:15 <denix> wasutton3_work: cross toolchain is NOT being used from local path
Jan 11 21:55:58 <wasutton3_work> hmmm. the build worked with yocto morty, so i guess ill diff the local.conf and see what changed?
Jan 11 21:59:05 <wasutton3_work> hmmm build seems to have resumed with removing one of the question marks after MACHINE in the local.conf
Jan 11 22:03:49 <wasutton3_work> hmmm same failure in the same package
Jan 11 23:50:46 <halstead> moto-timo, Does the oxygen eclipse plugin work with pyro? Asking for the yocto-2.3.3 release.
Jan 11 23:57:56 <halstead> moto-timo, There is an oxygen/pyro branch so I would guess yes but that might be there just to keep the build system happy.
Jan 12 00:37:17 <sum1> is the kernel maintainer back from vacation?
Jan 12 02:44:20 <sum1> zeddii: sup
**** ENDING LOGGING AT Fri Jan 12 03:00:02 2018