**** BEGIN LOGGING AT Fri Jun 29 03:00:00 2018
Jun 29 06:54:38 <yocti> New news from stackoverflow: Which Yocto branch to use <https://stackoverflow.com/questions/41632361/which-yocto-branch-to-use>
Jun 29 07:18:30 <learningc> I'm getting a compile error with qtbase-native.  How can I fix it?
Jun 29 07:26:06 <nayfe> learningc> can you pastebin it?
Jun 29 07:29:00 <learningc> There: https://pastebin.com/sXpQ39Kx
Jun 29 07:36:59 <nayfe> learningc did you try to clean it and rebuild?
Jun 29 07:38:22 <learningc> Yes
Jun 29 07:38:50 <learningc> Is there something that prevents it from building?
Jun 29 07:39:28 <learningc> When I rebuild it still cannot get past this error
Jun 29 07:51:44 <nayfe> "cc1plus: error: one or more PCH files were found, but they were invalid"
Jun 29 07:52:15 <nayfe> can you find a file named .pch/Qt5Concurrent in tmp build folder?
Jun 29 07:59:36 <nayfe> maxin: i replaced jamvm-native by cacao-native in java-native provider and it does not crash anymore, not sure what happens with jamvm ... I saw this patch https://github.com/PLi-metas/openembedded/blob/master/recipes/jamvm/files/debian-jni.patch but I'm not sure it applies on version jamvm_git (2.0)
Jun 29 08:00:16 <nayfe> maxin: btw hi again :)
Jun 29 08:17:03 <maxin> nayfe: good to hear that :)
Jun 29 09:43:32 <pouet_forever> Hey, I am trying to build the SDK with populate_sdk but I've got an issue...
Jun 29 09:44:03 <pouet_forever> I am running Debian 9 with gcc 4.9.2 and yocto 1.8
Jun 29 09:44:27 <pouet_forever> The error is : cat: /home/nico/yocto/ipos_v1.4/build/arago-tmp-glibc/work-shared/gcc-4.9.2-r0/gcc-4.9.2/gcc/defaults.h: No such file or directory
Jun 29 09:44:36 <pouet_forever> Any ideas ?
Jun 29 10:33:52 <bachp> pouet_forever: I found this in the mailing list: http://lists.openembedded.org/pipermail/openembedded-core/2015-April/103497.html seems to be a similar error. Could be related to your distro (arago).
Jun 29 10:39:33 <suvirb> how to update system time in yocto. it shows old date. root@nitrogen6x:~# date Thu Jun 22 23:56:28 UTC 2017.
Jun 29 10:39:41 <suvirb> even though i'm connected to internet
Jun 29 11:08:16 <pouet_forever> bachp: thank you ! I have already found this post before but there is no solution to solve my problem
Jun 29 11:09:21 <pouet_forever> I have extracted the gcc-linaro-4.9-2015.03.tar.xz file and the file defaults.h is present... :-/
Jun 29 11:55:39 <yocti> New news from stackoverflow: How to get bridge support with /etc/network/interfaces? <https://stackoverflow.com/questions/51100710/how-to-get-bridge-support-with-etc-network-interfaces>
Jun 29 12:05:43 <steeno> Hi, one short question. Old rpi projects based on jethro got migrated to morty, some package architectures changed now from cortexa7hf-vfp-vfpv4-neon to cortexa7hf-neon-vfpv4, therefore new package versions get not recognized via opkg.
Jun 29 12:06:30 <steeno> is there a possibility to change the target architecture back to the old wording?
Jun 29 12:09:34 <steeno> i found the package_extra_archs var in the arm in machine/include/tune-cortexa7.inc f.e. but i don't really understand the system which determines what package is valid for which cpu arch
Jun 29 12:25:45 <yocti> New news from stackoverflow: How to install dlib and update gcc on Yocto <https://stackoverflow.com/questions/47666774/how-to-install-dlib-and-update-gcc-on-yocto>
Jun 29 12:36:25 <yates> what does "do_install" do? we're building an image, not installing a package, so what is being installed to where? --Confused
Jun 29 12:38:04 <eduardas_m> hello, I believe I can see a problem with the bmap-tools recipe in master... python3-fcntl is not listed as a dependency, though it is needed to import during runtime
Jun 29 12:38:15 <fsdun> yates, the software needs to be installed into a pseudo root to be splitted into it's packages
Jun 29 12:39:28 <yates> fsdun: thanks - makes sense now.
Jun 29 12:39:55 <fsdun> eduardas_m, will it be pulled in through RDEPENDS packages?
Jun 29 12:40:19 <eduardas_m> fsdun: no, that is the point
Jun 29 12:41:18 <eduardas_m> fsdun: http://cgit.openembedded.org/openembedded-core/tree/meta/recipes-support/bmap-tools/bmap-tools_3.4.bb?h=master
Jun 29 12:41:57 <eduardas_m> fsdun: python3-fcntl is not listed as a direct RDEPENDS and neither does it get pulled in indirectly
Jun 29 12:42:50 <eduardas_m> the only reason it used to work for me is becaused it used to be pulled in by unrelated packages, as soon as I slimmed down my image, it broke
Jun 29 12:42:52 <fsdun> In this case you should send a fix to ML
Jun 29 12:48:12 <yates> fsdun: where is this package-building done? is it somewhere under the tmp/ folder?
Jun 29 12:48:35 <yates> rather, package-splitting
Jun 29 12:50:15 <yates> i'm getting this error: "...libubi.h': No such file or directory" https://paste.fedoraproject.org/paste/82zN5meLNoWDlV0M1o9R8Q
Jun 29 12:59:55 <yates> nm, i found it.
Jun 29 13:25:57 <yocti> New news from stackoverflow: compile Dlib library with Yocto <https://stackoverflow.com/questions/51102780/compile-dlib-library-with-yocto>
Jun 29 13:38:30 <yates> where is mkdep defined? $(foreach v,$(MTD_BINS),$(eval $(call mkdep,,$(v))))
Jun 29 13:38:46 <yates> is it built-in to gnumake?
Jun 29 14:37:14 <yates> nm found it
Jun 29 15:40:43 <zeddii_home> when I’m on vacation, it would be easier if everyone else stopped working as well! :P
Jun 29 16:58:49 <armpit> RP, when does 'LAYERSERIES_CORENAMES" change?
Jun 29 17:00:46 <RP> armpit: I need to add the next one now
Jun 29 17:01:01 <RP> armpit: then we drop sumo when we break 'ABI'
Jun 29 17:55:21 <mihai> is there a lag on the mailing lists?
Jun 29 17:56:13 <mihai> or just my email getting stuck somewhere along the way
Jun 29 17:56:16 <JPEW> I just sent an email and it went through right away
Jun 29 17:56:33 <JPEW> If you're not signed up for the list I think the messages are moderated?
Jun 29 17:59:22 <mihai> I should be signed up
Jun 29 18:00:24 <mihai> I have the welcome email
Jun 29 18:00:38 <mihai> maybe there's a trial period :)
Jun 29 18:01:14 <yates> when bitbake runs a gnumake command as part of a do_install for a recipe, how does it form the make command, e.g., how are -I options constructed?
Jun 29 18:02:25 <yates> no, wait, the compiler -I options are in the makefile
Jun 29 18:10:37 <yates> how can the install command complain about libubi.h not being found if that is not in the list of input files it is to copy? https://paste.fedoraproject.org/paste/QUzYzDtR7L0-0inTckDyDQ
Jun 29 18:13:37 <yates> also, why is the "install: cannot stat ..." being emitted after the gzip command and not after the install command? is it to do with -j options for the makefile?
Jun 29 18:22:03 <JPEW> Is there maybe one of your layers that is modifying mtd-utils with a .bbappend?
Jun 29 18:23:00 <mihai> yates, not likely, that make seems to be on a single thread
Jun 29 18:23:12 <mihai> see /Storage/swupdate-test/sources/poky/build-hw-test-image/tmp/work/armv7at2hf-neon-fslc-linux-gnueabi/mtd-utils/1.5.2-r0/temp/run.do_install.18064
Jun 29 18:23:27 <mihai> maybe there's an install after that gzip
Jun 29 18:39:26 <yates> mihai: aha.
Jun 29 18:39:29 <yates> AHA!
Jun 29 18:41:23 <yates> well, wait
Jun 29 18:42:27 <yates> is this python? https://paste.fedoraproject.org/paste/kyK1jmCzkj8vJV1nultnwQ
Jun 29 18:42:50 <yates> no, bash script
Jun 29 18:48:59 <yates> i'm really confused. is this logfile output from the do_install() command in the shell script, or from the makefile?
Jun 29 18:49:42 <yates> the first line of do_install() is to run the makefile (oe_runmake... ), right?
Jun 29 18:50:15 <yates> ok, that's right.
Jun 29 18:50:51 <yates> yes, it runs the make file, the last command of the make install target being gzip, THEN it runs the install's after the oe_runmake.
Jun 29 18:52:35 <yates> where does this script come from? is it in the recipe somewhere? (some recipe?)
Jun 29 18:54:24 <kergoth> read the bbclasses the recipe inherits
Jun 29 18:54:38 <kergoth> including the globally inherited classes, 'base' and any listed in the INHERIT variable
Jun 29 18:54:46 <kergoth> but its easier to just use bitbake -e as was suggested yesterday
Jun 29 18:54:54 <kergoth> which includes the definitions of all the defined functions
Jun 29 18:54:56 <kergoth> including oe_runmake
Jun 29 19:11:48 <yates> kergoth: ok, i have this massive -e output. i'm not sure how to use it to find where this install script is coing from
Jun 29 19:12:05 <yates> coming
Jun 29 19:12:21 <kergoth> i don't understandt heq uestion
Jun 29 19:12:26 <kergoth> you don't know how to search a text file?
Jun 29 19:12:28 <kergoth> grep? vim?
Jun 29 19:12:38 <kergoth> look for oe_runmake () {
Jun 29 19:12:55 <kergoth> it's basically shell script, other than the python functions being included as well
Jun 29 19:13:07 <kergoth> if you can read a recipe, you can read the -e output
Jun 29 19:13:11 <kergoth> it's the same syntax
Jun 29 19:13:47 <yates> i'm looking for the do_install for that particular recipe/thing, and there are a bazillion do_installs
Jun 29 19:14:31 <kergoth> no, there isn't
Jun 29 19:14:40 <kergoth> bitbake -e somerecipe, search for do_install () {
Jun 29 19:15:03 <yates> there is only one do_install for the entire bitbake build?
Jun 29 19:15:05 <kergoth> not that hard to follow it. do_install () { } likely runs autotools_do_install, so then search for that
Jun 29 19:15:13 <kergoth> no, bitbake -e specifies the recipe
Jun 29 19:15:20 <kergoth> as you could tell if you'd run bitbake --help
Jun 29 19:15:27 <kergoth> bitbake -e somerecipe shows the metadata for that recipe
Jun 29 19:15:39 <kergoth> bitbake -e is just the global configuration metadata (bitbake.conf, local.conf, etc)
Jun 29 19:15:44 <kergoth> (not specifying the recipe)
Jun 29 19:16:39 <kergoth> no, there aren't
Jun 29 19:16:47 <kergoth> erm ,wrong window on that
Jun 29 19:16:58 <kergoth> exactly what problem are you hitting? i don't have the full scrollback from this irc client
Jun 29 19:19:28 <kergoth> hmm
Jun 29 19:20:11 <yates> "bitbake swupdate-image" errors out with "cannot stat libubi.h". that was traced to the do_install() function in a bash shell script via the run.do_install.xyz file, specifically to an "install" command specifying the location of the libubi.h i want to find out where that do_install script is generated
Jun 29 19:20:34 <yates> "do_install ()" is not found in the -e output
Jun 29 19:20:45 <kergoth> if do_install wasn't in -e, then tehre is no do_install
Jun 29 19:20:53 <kergoth> -e dumps the full metadata
Jun 29 19:20:57 <rburton> so swupdate-image is an image, and doesn't do any building
Jun 29 19:21:11 <kergoth> yeah, as rburton says, you need to examine the metadata of the recipe that's failing, with -e, not the image
Jun 29 19:21:12 <rburton> you want to read the log to find out what dependent recipe was being built, and look at *that* recipe
Jun 29 19:23:21 <rburton> the recipe most likely doesn't have a do_install directly but uses the default from base.bbclass which essentially just does 'make install', so you need to look at the makefiles
Jun 29 19:23:33 <rburton> but worth checking the recipe doesn't have manual install steps
Jun 29 19:24:39 <yates> rburton: that is not the case here. there is a 'make install' as well as several other commands e.g., 'install' in a bash script.
Jun 29 19:24:58 <yates> and the error is coming from one of the script 'install's
Jun 29 19:25:14 <rburton> down the rabbit hole you go, welcome to the wonderful world of integration
Jun 29 19:25:24 <yates> ha.
Jun 29 19:25:33 <rburton> this week i fixed a bug in a configure script which uses sh to generate makefile rules
Jun 29 19:25:36 <yates> did i take the red pill or the blue pill?
Jun 29 19:26:00 <yates> sounds horrendous
Jun 29 19:27:14 <rburton> yeah it wasn't fun, and left me wondering who in their right mind would construct a build system like that
Jun 29 19:27:21 <rburton> (the libsdl maintainers, is the answer)
Jun 29 19:28:14 <yates> ok, i may get booted from teh channel (or /ignored..), but can you give me a clue: https://paste.fedoraproject.org/paste/eNU4UMdFxPB73HcymvG4Gg
Jun 29 19:28:53 <yates> how do i look through this to see where this do_install shell script is generated?!?
Jun 29 19:30:05 <rburton> presumably thats the wrong recipe still, that has a do_install that calls base_do_install that does nothing
Jun 29 19:30:38 <rburton> PN="swupdate-image" <-- you're still doing -e on the image, not the recipe that breaks
Jun 29 19:30:54 <yates> rburton: when you said "15:21 <rburton> you want to read the log to find out what dependent ..." did you mean the output of -e for the image?
Jun 29 19:31:12 <rburton> no
Jun 29 19:31:13 <otavio[m]> armpit:
Jun 29 19:31:17 <rburton> you run bitbake swupdate-image
Jun 29 19:31:19 <rburton> it breaks
Jun 29 19:31:25 <yates> right
Jun 29 19:31:29 <rburton> the log on screen tells you what recipe broke
Jun 29 19:31:34 <otavio[m]> armpit: how are you?
Jun 29 19:31:36 <armpit> otavio[m], yes
Jun 29 19:31:40 <armpit> fine
Jun 29 19:31:47 <otavio[m]> could you review https://github.com/OSSystems/oe-core/compare/sumo...sumo-next for stable?
Jun 29 19:32:37 <yates> ok
Jun 29 19:33:21 <yates> so you mean this: ERROR: Task (/Storage/swupdate-test/sources/poky/meta/recipes-devtools/mtd/mtd-utils_git.bb:do_install) failed with exit code '1'
Jun 29 19:33:31 <armpit> otavio[m], sure thing.. thanks
Jun 29 19:33:41 <yates> so poky/meta/recipes-devtools/mtd/mtd-utils_git.bb broke
Jun 29 19:34:07 <JPEW> yates: Yes, you want 'bitbake -e mtd-utils'
Jun 29 19:34:07 <yates> so then i should run bitbake -e mtd-utils_git ?
Jun 29 19:34:30 <JPEW> _git is the "version", so you omit it
Jun 29 19:37:10 <rburton> yates: easier to just look at the recipe tbh
Jun 29 19:37:31 <rburton> see if it has a do_install itself, if not look at the classes it inherits,
Jun 29 19:38:08 <yates> ok, i found the do_install in the -e output with the offending "install". yay!
Jun 29 19:38:31 <yates> it's "do_install()", btw. no space before the "()"
Jun 29 19:39:07 <yates> rburton: ok.
Jun 29 19:41:25 <yates> there is a "do_install () {" function defined in poky/meta/recipes-devtools/mtd/mtd-utils_git.bb, but it doesn't match the shell script i saw.
Jun 29 19:41:39 <yates> it only has the first oe_runmake: https://paste.fedoraproject.org/paste/x2lZx-jq7M-LQzEdWw~~bA
Jun 29 19:42:11 <rburton> where is this script you saw?
Jun 29 19:42:26 <rburton> sounds like its from the makefile to me
Jun 29 19:42:35 <rburton> as you see, the recipe just does 'make install'
Jun 29 19:42:54 <yates> it was in poky/build-hw-test-image/tmp/work/armv7at2hf-neon-fslc-linux-gnueabi/mtd-utils/1.5.2-r0/temp/run.do_install.18064
Jun 29 19:43:00 <yates> https://paste.fedoraproject.org/paste/H~Z5G6N5qQCSfSnskOgmUA
Jun 29 19:45:02 <rburton> looks a lot like you've inherited a class globally that really shouldn't be inherited globally
Jun 29 19:45:43 <rburton> whatever that is, its part of swupdate, so i'd be asking the maintainers of that what broke
Jun 29 19:45:54 <yates> babic
Jun 29 19:45:59 <yates> stefano babic
Jun 29 20:33:23 <yates> ok, posted on https://groups.google.com/forum/#!forum/swupdate
**** ENDING LOGGING AT Sat Jun 30 03:00:03 2018