**** BEGIN LOGGING AT Fri Apr 28 03:00:01 2017 Apr 28 09:29:03 khem: happened again https://autobuilder.yocto.io/builders/nightly-mips64/builds/246 Apr 28 09:40:08 urg Apr 28 09:40:28 so I have a tarball untarring and creating a file owned by root Apr 28 09:40:35 only happens on one machine Apr 28 09:41:56 * Crofton|work will now go insane Apr 28 09:42:26 popcorn? Apr 28 09:42:59 enoy watching me burn Apr 28 09:51:44 will do! *nomnomnom* Apr 28 09:57:16 fark, need root pw to get docker to run Apr 28 09:58:54 yeah, common misconception is that docker is about security, respectively containment, which it is not. docker is at its heart only about namespaces - so if you have docker access, you effectively are root. Apr 28 09:59:05 yeah Apr 28 09:59:08 this is fine with me Apr 28 09:59:32 just annoying the guy that set this up didn't make sudo nopasswd :) Apr 28 09:59:40 hrhr Apr 28 10:00:18 this is just to isolate me from host version Apr 28 10:00:36 apparently there are still holes in my cunning plan :) Apr 28 10:01:27 need some polyfill? Apr 28 10:02:23 I need a clever intern who knows all this modern stuff Apr 28 10:02:45 and then give him the mailaddr polyfill@... Apr 28 10:07:51 rburton: would it make sense to move nios2 tunes from meta-altera to oe-core ? Apr 28 10:08:11 rburton: with that, oe-core would have nios2 support Apr 28 10:12:22 could replace mips Apr 28 10:12:29 with that altera crap Apr 28 10:13:32 anyone know of a bug in tar that gets the file persm wrong on the last file in a tarball? Apr 28 10:16:06 * Marex smacks Crofton|work Apr 28 10:16:27 wouldn't it make more sense to have all the arch bits in core ? Apr 28 10:16:35 board bits can be in meta-foo layers Apr 28 10:19:25 one in one out, drop mips and we can add altera ;) Apr 28 10:19:35 (that's not the policy really) Apr 28 10:20:52 rburton: I figured that much, heh Apr 28 10:21:00 rburton: well, why dropping the mips ? Apr 28 10:21:08 rburton: what are you getting at ? Apr 28 10:21:13 does anyone use mips? :) Apr 28 10:22:08 rburton: yes, me :) Apr 28 10:22:34 Marex: oh good, you can help khem fix gcc7 on mips then, it breaks :) Apr 28 10:23:02 ah shit, that was low Apr 28 10:23:06 rburton: what's the problem ? Apr 28 10:24:59 :) Apr 28 10:25:47 Marex: not sure beyond "autobuilder fails on mips-lsb and mips64". they claim the qemu doesn't boot. i haven't tried myself to see what happens yet. Apr 28 10:28:55 rburton: well, my crystal ball needs to be recharged ... they don't make them like they used to with all the fancy newfangled 3D crap and animations and slim body nowadays ... Apr 28 10:38:56 Marex: if you add nios to oe, then we can add microblaze too :P Apr 28 10:40:27 nrossi: something something altera > xilinx something :) Apr 28 10:41:11 Marex: dont you mean, "something something intel > xilinx something something complete" ? :D Apr 28 10:41:59 nrossi: I think this discussion is officially something something complete :) Apr 28 10:42:31 nrossi: well why don't we add microblaze too anyway ? Apr 28 10:42:54 Marex: im for it, just never had a reason to do it is all Apr 28 10:43:52 and thus my question, does it make sense to keep all the architecture bits in meta- or would it make more sense to move them to core and keep only the board stuff in meta-foo ? Apr 28 10:44:03 arch bits are in oe-core if they can be tested in oe-core Apr 28 10:44:30 so add a qemunios machine and we can talk Apr 28 10:44:34 rburton: super, so once I update qemu to 2.9.0 , I can get nios2 into oe-core Apr 28 10:45:08 rburton: nios2 landed in 2.9.0 , so that's perfect timing Apr 28 10:46:21 rburton: if gcc/binutils/glibc patches are required would they be acceptable in oe-core? (microblaze still needs a few for full functionality) Apr 28 10:46:58 i can see merit in merging those Apr 28 10:47:23 nrossi: well that's a bad bad xilinx :-) iirc I don't have any for nios2 Apr 28 10:47:38 Marex: yes that is a bad bad xilinx... :P Apr 28 10:48:14 rburton: well that sounds like a plan then, will sort out getting that stuff in the works for master :) Apr 28 10:49:49 sounds great indeed Apr 28 10:50:03 rburton: thanks Apr 28 13:32:22 Today's riddle: http://paste.debian.net/929869/ Apr 28 13:33:02 jethro in a Fedora 22 container on Ubuntu Apr 28 13:33:16 I do not see this in same container on Fedora 25 Apr 28 13:35:29 Crofton|work: a docker bug that's been fixed post-F22? Apr 28 13:36:21 I wonder if this ubuntu machine is up to date? Apr 28 13:36:33 Fedora 22 on Fedora 25 host is fine Apr 28 13:36:42 Fedora 22 on some Ubuntu isn't Apr 28 13:37:12 yes, need to check if this is up to date Apr 28 13:46:35 Hello. I have been trying to build a toolchain for Qt5 devlopment from my image which built successfuly. The tooolchain build keeps failing on errors regarding QtScript. I went into the log file but I cannot make much sense of it to come to a solution. Apr 28 13:46:43 This is my log file output https://pastebin.com/zm5wtEL4 Apr 28 13:47:27 Has anyone run into this issue or have a possible solution? I never even included qtscript in my local.conf Apr 28 15:12:40 smurray, so updated ubuntu 16.04 LTS and still have wierd issue with some files ending up in work owned by root Apr 28 15:13:01 I'm going to start drinking Apr 28 15:13:51 Crofton|work: use kvm to run F25 on the Ubuntu, then run your container inside that? ;) Apr 28 15:14:02 lol Apr 28 15:14:13 this is the sort of crap that shouldn't happen Apr 28 15:14:29 it blow my mind how the file gets created as root in the first place Apr 28 15:17:28 somewhat concerning isn't it Apr 28 15:20:08 Crofton|work: use craps to dockerify your yocto^WOE build ? Apr 28 15:20:23 it is dockerized Apr 28 15:20:30 rburton, yes Apr 28 15:20:37 * Crofton|work blames ubuntu Apr 28 15:20:50 * Marex suspects PEBKAC or docker issue Apr 28 15:21:14 lol Apr 28 15:21:16 Crofton|work: hey you aren't running docker with overlayfs or backend, are you ? Apr 28 15:21:45 :~/Workarea/oe-jethro$ sudo docker run -i -t --rm -v $(pwd):$(pwd) oe-jethro:22 /bin/bash Apr 28 15:21:47 Crofton|work: they're using different storage backend on debian and fedora, I ran into it at one point Apr 28 15:22:15 * Crofton|work is an idiot about this stuff, he just googles and copies and fiddles till things work Apr 28 15:22:59 * Marex believes Crofton|work ... totally Apr 28 15:24:13 ok, back to making embedded zoo great again Apr 28 15:24:37 Crofton|work: -v foo:bar bind-mounts the dir from host into the container ? Apr 28 15:24:57 Crofton|work: arent you running as root in the container ? Apr 28 15:25:19 not tunning as root Apr 28 15:25:21 hmmm Apr 28 15:25:49 Crofton|work, did you stay awake all night or woke early to do this mess? Apr 28 15:25:58 woke early Apr 28 15:26:13 I see beer time here Apr 28 15:26:46 I'm following your struggle since this morning and was a bit concerned Apr 28 15:27:40 fwiw I had an allucination yesterday night: the images in deploydir had different permissions... Apr 28 15:28:12 in the left pane musl, in the right pane glibc,...I'll check if some are possessed by root :) Apr 28 15:28:25 demonic possesions Apr 28 15:29:39 ant_work: how's my MTD parser victim doing ? Apr 28 15:29:43 Crofton|work, strangest things happen these days: I'm pretty sure the first build produced bigger mtd images than the second... Apr 28 15:30:20 Marex, lazy cow, do pls suggest improvements ;) Apr 28 15:31:19 great Apr 28 15:31:25 * Marex rebuilds his glorious nios2 toolchain Apr 28 15:31:29 the wierd thing is the owner id correct "most" of the time Apr 28 15:32:20 Crofton|work: it's haunted :) Apr 28 15:38:26 Crofton|work, does it happen only in WORKDIR? Apr 28 15:39:19 I don't think so, just looking thre Apr 28 15:39:56 I build on Ubuntu 16.04 at home, I can verify. I build in tmpfs however, with mode=1777 Apr 28 15:40:30 I'll double check DEPLOYDIR Apr 28 15:46:54 ant_work, are you running a Fedora container on Ubuntu? Apr 28 15:47:11 no, sorry Apr 28 15:48:30 Marex, khem was hinting at https://www.openembedded.org/wiki/Commit_Patch_Message_Guidelines Apr 28 15:49:34 ant_work: so instead of fixing minor detail up and applying a bugfix, we'll leave the bug unfixed ... good ! Apr 28 15:49:53 Marex, no, we aren't kernel.org here ;) Apr 28 15:50:20 *shrug* Apr 28 15:50:41 http://cgit.openembedded.org/meta-openembedded/log/?h=master-next&qt=grep&q=fbida Apr 28 15:50:48 is in master-next since some days Apr 28 15:52:00 super Apr 28 15:52:03 then it's done Apr 28 15:52:20 * Marex goes back to making embedded zoo great again Apr 28 15:52:45 now is merge-window again in OE Apr 28 15:52:49 ;) Apr 28 18:11:54 rburton: ok, can someone readout possible error that might be happening Apr 28 18:12:12 rburton: I have pushed grub-2.02 update to kraj/master Apr 28 18:12:31 rburton: also dropped git recipe since its not useful any more Apr 28 19:26:49 khem: sweet **** BEGIN LOGGING AT Sat Apr 29 00:04:53 2017 **** ENDING LOGGING AT Sat Apr 29 03:00:01 2017