**** BEGIN LOGGING AT Mon Apr 30 03:00:06 2018 Apr 30 06:28:07 any one launching qemu for qemuarmarm 64 Apr 30 06:29:33 while launching it is displaying as kernel panic: not sincing- invalid configuration from end user Apr 30 06:30:03 any help please to overcome it Apr 30 06:44:16 New news from stackoverflow: Yocto : How To build Image file for Banana Pi M1+ with the help of Yocto Project Apr 30 07:14:22 New news from stackoverflow: How to change the config of u-boot in Yocto Apr 30 09:44:49 New news from stackoverflow: Copy Yocto Project to other PC by tar || In yocto embedd system for source env/bin/activate : unable to execute arm-poky-linux-gnueabi-gcc Apr 30 15:15:50 New news from stackoverflow: How to Enable or configure ftrace module Apr 30 15:23:04 rburton: yt ? Apr 30 15:23:58 rburton: Drop all my changes and pick http://git.openembedded.org/openembedded-core-contrib/log/?h=kraj/pu Apr 30 15:24:31 khem: trying to test isolated bits as if it breaks with 25 patches then it can be a pain to bisect Apr 30 15:25:39 I tested them together Apr 30 15:26:04 and the ab is telling me that build-applicance is failing, there's warnings from multiple providers, and that musl is runtime failing Apr 30 15:26:31 now i'm kind of pleased that i can blame just five patches instead of 25 when i tell you that musl images build but don't work Apr 30 15:28:42 may be drop musl upgrade to test the rest http://git.openembedded.org/openembedded-core-contrib/commit/?h=kraj/pu&id=e6ffc6dbe7f979bbc2fa11dcfd72b9b22c4222f8 Apr 30 15:28:52 oh i see the packagegroup change has more changes than the shortlog suggests Apr 30 15:29:06 yes running a test build now without the musl upgrade Apr 30 15:29:35 the gcc-runtime patch has changed too from last week Apr 30 15:30:02 since I got some cycles for building mingw32 I found that I was using wrong overrides Apr 30 15:31:29 whats the error regarding multiple providers Apr 30 15:34:05 NOTE: Multiple providers are available for virtual/crypt (glibc, musl) Apr 30 15:35:12 must be related to a specific configuration as the relevant change is in master but i've never seen it Apr 30 15:36:16 khem: whilst i've got you, does llvm inherit pythonnative and perlnative only because cmake is dumb and not because it actually needs them Apr 30 15:36:24 got a one-liner to remove pythonnative already Apr 30 15:45:58 New news from stackoverflow: write systemd script for influxdb for a yocto build Apr 30 16:03:24 rburton: it needed them Apr 30 16:04:07 and some hosts have different opinion on python if it should be py3 or py2 Apr 30 16:04:25 so it insulates the build from those kind of issues Apr 30 16:06:22 rburton: virtual/crypt does this show up with musl builds Apr 30 16:06:42 it was the buildappliance build iirc Apr 30 16:07:21 so that must be glibc Apr 30 16:09:59 bitbake build-appliance-image doesnt show this here Apr 30 16:14:31 even with TCLIBC=musl it doesnt show up Apr 30 16:57:07 having an issue with opkg not running preinst scripts when using offline-root Apr 30 16:57:44 i found the init script to trigger postinst on reboot but don't see anything to trigger preinst Apr 30 17:00:23 this is when upgrading an existing system. they run during image creation. Apr 30 17:00:37 If ${THISDIR} is used on a recipe and it's bbappend file located in different layers, should the value be same across different layers? Apr 30 17:01:25 Guest76094: there's a reason FILESEXTRAPATHS_prepend uses := Apr 30 17:01:35 forces immediate expansion, so it uses the file currently being parsed, rather than the recipe itself Apr 30 17:04:00 i have a problem where do_install doesn't pick files from /files/ although bbappend file contains FILESEXTRAPATHS_prepend := "${THISDIR}/files:" Apr 30 17:04:38 and when i do a echo ${THISDIR} from .bbappend, it prints the path where the .bb file is located. Apr 30 17:05:22 of course, because the ${THISDIR} you used in do_install was expanded later, not immediately the way the reference in FILESEXTRAPATHS was Apr 30 17:05:29 use bitbake -e to examine FILESEXTRAPATHS instead Apr 30 17:08:19 but shouldn't the files be fetched to WORKDIR with EXTERNALSRC build? Apr 30 17:08:40 you said nothing about externalsrc Apr 30 17:08:48 externalsrc builds don't unpack at all, by definition Apr 30 17:08:52 sorry my bas Apr 30 17:08:52 d Apr 30 17:08:57 file:// files are unpacked to workdir in do_unpack Apr 30 17:09:15 externalsrc == use existing sources, don't fetch, unpack, or patch Apr 30 17:10:12 in that case, we should use THISDIR for installing file:// URLS if i am not wrong. Apr 30 17:11:13 i don't understand what you're saying. we already use thisdir for installing file:// urls, that's what FILESEXTRAPATHS *is* Apr 30 17:14:32 ok, i could see FILESEXTRAPATHS contains the path from where i am trying to install, but do_install complains about the non-existing file in the path where .bb file is hosted! Apr 30 17:14:52 am i missing anything? Apr 30 17:59:55 khem: Looks like some sanity test failures on the ross/ssp branch Apr 30 18:00:06 khem: https://wiki.yoctoproject.org/wiki/BuildLog#nightly_1061_-_ross.2Fssp_8a17b60c460f2662e9308f68355db71d4ca4a6f3 Apr 30 18:00:13 khem: just a head's up :) Apr 30 18:03:37 stephano: yeah ssh: connect to host 192.168.7.4 port 22: Connection refused Apr 30 18:03:53 thats musl things, I test on rpi3 and it works alright Apr 30 18:04:12 the appliance issue is more related to ross not picking right bits Apr 30 18:04:20 ha ha ha, yes Apr 30 18:04:29 okay, good, just being sure you knew Apr 30 18:05:00 are these images using systemd ? Apr 30 18:05:59 I believe so... but I'd want to check to be sure Apr 30 18:08:09 seems not, but they use openssh and I use dropbear Apr 30 18:09:04 khem: checking now Apr 30 18:09:06 and I have seen openssh not let a connection go through from my arch box Apr 30 18:09:20 probably some mismatches in client version Apr 30 18:09:55 :Q Apr 30 18:09:58 :) Apr 30 18:10:06 * stephano finds the correct window Apr 30 18:12:51 hmm, that's on tumbleweed, so we can just ignore any errors, right? :-D Apr 30 18:13:33 tumbleweed has weed in its name Apr 30 18:14:23 hmm oe-core/master is using linux-yocto 4.8 for qemux86 is that intentional ? Apr 30 18:14:32 poky seems to be using 4.14 Apr 30 18:15:45 you're right, no systemd Apr 30 18:18:13 * stephano --> lunch (brb) Apr 30 18:55:47 * armpit great.. same commit new build failure Apr 30 19:10:18 khem: oe-core should be using the lastest Apr 30 19:10:40 in fact, should it ever pin a version? Apr 30 23:00:07 armpit: should I be merging some stable patches? If so let me know which ones are ready... Apr 30 23:01:24 RP, rocko-next passed last week, failed today with no changes.. Apr 30 23:01:56 I started to separate the good last build from what I am collecting Apr 30 23:02:15 just to keep me sane ( if that is possible) Apr 30 23:03:49 * armpit stable/*-mnut Apr 30 23:23:30 ~curse rocko-next Apr 30 23:25:00 * tlwoerner checks users list ... :-( May 01 00:38:24 tlwoerner, why May 01 01:14:03 should I be sending oe-core recipe updates (incremental) against master or master-next? or wait until sumo is out and master-next merges to master? **** ENDING LOGGING AT Tue May 01 03:00:05 2018