**** BEGIN LOGGING AT Tue Dec 04 03:00:16 2018 Dec 04 03:21:23 shutil.move() seems to work but puts all the files in a deeper subdirectory than it should Dec 04 03:21:54 and if it takes me another 45 minute runtime to find out if trying a different function works or not, then this is definitely my last attempt. Dec 04 03:27:29 looks like it will.. every time i make a change to one of the scripts, it has to re-run a couple hundred tasks :| Dec 04 05:48:19 :O i think i fixed it. made it shutil.rmtree() then shutil.copytree() ... at least, it passed one of the failing packages without errors, and the packages directory looks correct now Dec 04 07:57:15 Tofe: Herrie|2: I'm going to debug qemux86 kernel build directly on jenkins, please don't trigger any builds for a while Dec 04 08:10:35 JaMa: ok ! Dec 04 08:10:47 Good luck Dec 04 08:14:59 Morning Dec 04 08:15:03 Tofe: Checking your email Dec 04 08:19:05 other useful information: if for some reason (such as you've never made a commit on a machine) you have not configured your git user identity, a LuneOS build may fail due to the android-tools package erroring out Dec 04 08:19:48 | fatal: empty ident name (for ) not allowed Dec 04 08:20:18 EricBlade: Ah that's good to know Dec 04 08:20:49 Tofe: Herrie|Laptop: done, jenkins is available again, fix merged in pyro Dec 04 08:21:27 JaMa: that was fast :) Dec 04 08:29:14 EricBlade: Your WSL fixes might be good to send upstream Dec 04 08:29:42 I doubt many people @ Yocto and Intel care about it too much, but if it's conditional and doesn't break stuff on Linux why not? Dec 04 08:29:48 JaMa: ^ thoughts on this? Dec 04 08:30:48 i'm not positive that this is a solution just yet, and i haven't used Python in *so* long i have to look up the syntax for everything. :-D Dec 04 08:32:39 there's a weird problem that seems to be occurring in the base-files class, where it seems that a file in the build directory gets deleted somehow while the package step is running. i could be wrong, though, maybe the file never actually exists. and strace on processes like bitbake are a nightmare Dec 04 08:32:59 s/base-files class/base-files recipe/ Dec 04 08:34:23 so what i've got right now is not at all conditional, *and* it's really doing different operations, hoping for the same outcome at the end. Dec 04 08:42:09 latest oe-core explicitly checks for WSL and refuses to pass sanity checks, because WSL is not ready Dec 04 08:42:37 http://git.openembedded.org/openembedded-core/commit/?id=62fe04c2451f612b9ee5eba469c7724b63484499 Dec 04 08:43:56 if it all works, then upstream will be interested, but it's WSL filesystem limitation which will be hard to work around in bitbake, so I guess someone needs to fix WSL first Dec 04 08:44:52 JaMa: EricBlade is trying to work his way through and fix workarounds/fixes Dec 04 08:45:33 having it restart from scratch every time i change a class is a bit daunting. :-D Dec 04 08:45:43 not sure if it's worth it as krogoth said: Even if bitbake could run on WSL, it’d be horribly slow. I/O on WSL is *terrible* right now, docker or a VM wipes the floor with it. I’m sure it’ll be viable eventually, though :) Dec 04 08:46:34 OH, this does remind me of something I did want to bring up to you, JaMa.. not sure if it applies to current oe-core, but i noticed that using a git submodule checkout as a build directory fails Dec 04 08:46:35 http://lists.openembedded.org/pipermail/openembedded-core/2018-May/150635.html Dec 04 08:46:56 EricBlade1: which version of oe-core? Dec 04 08:47:17 EricBlade1: gitsm fetcher was reworked late in thud, some fixes weren't backported from master to thud yet Dec 04 08:48:07 if you can reproduce the issue with latest bitbake, then write an e-mail to bitbake-devel and cc Mark Hatle who did the rework Dec 04 08:49:13 good question. i don't know how to find the answer on that. somewhere in one of the core classes, it does some operation on a .git directory if you're building with externalsrc Dec 04 08:49:32 but in a submodule checkout, .git is a file Dec 04 08:55:12 thanks for the pointer, JaMa. looks like problem has been solved at some point. srctree_hash_files execs git to ask it where the git dir is, if any, instead of assuming it's ".git" in the current directory. Dec 04 09:08:46 novaldex: ping Dec 04 09:13:34 Herrie|Laptop: pong, good morning Dec 04 09:13:49 novaldex: Morning! Dec 04 09:13:56 Any ideas what's up with issues.webos-ports.org ? Dec 04 09:17:00 wasn't aware there were any! let me have a peek Dec 04 09:18:03 hmmm, looks like DNS, okay i'm on it Dec 04 09:19:23 novaldex: Thnx Dec 04 09:19:30 And for the Preware feed, any clues there? Dec 04 09:19:50 I don't see anything obvious. Both LuneOS and WebOS Internals feeds are doing exactly the same Dec 04 09:20:01 Both running on Bonaire, just pointing to a different repo Dec 04 09:20:09 Only thing I can think of is a missing key somehow Dec 04 10:11:16 Jay finally can reach my home PC with VNC Dec 04 10:12:08 Turned out I needed to change a setting in my Lyra mesh, by default they were acting as router instead of AP Dec 04 11:04:23 seems to be an issue with the VM that hosts the issues site, will take me a bit longer to look into it Dec 04 11:11:27 novaldex: Thnx Dec 04 11:11:40 In general this ChiliProject should be migrated sometime as well Dec 04 11:16:39 I guess we could go back to RedMine or OpenProject. Dec 04 11:16:57 Seems OpenProject is a bit more actively developed.... https://www.openproject.org/operations/upgrading/migration-path-from-chilliproject-to-openproject/ Dec 04 11:26:23 Migration seems pretty straight forward Dec 04 13:33:22 Tofe: now the only failure in testing builds is nyx-modules for tenderloin, the file is missing in your fork we're currently using, was it intentional? Dec 04 13:33:25 Tofe: nyx-modules $ git diff wp/webOS-ports/master tofee/tofe/webosose --stat -- files/systemd/nyx.target.d/wait-touchscreen.conf Dec 04 13:33:28 files/systemd/nyx.target.d/wait-touchscreen.conf | 3 --- Dec 04 13:35:58 JaMa: I'd say it's oversight and we'd need to add it back Dec 04 13:37:13 I was wondering from where it is being installed and install() is added through meta-luneos/recipes-webos/nyx-modules/nyx-modules/tenderloin.cmake Dec 04 13:37:47 It's a bit hacky solution imho, but I'm sure nizovn had a good reason to add it like this Dec 04 13:37:49 He usually has Dec 04 13:39:54 and the path where it's included is also slightly different: Dec 04 13:39:56 - mkdir -p ${S}/src/machine Dec 04 13:39:56 - cp ${WORKDIR}/${MACHINE}.cmake ${S}/src/machine Dec 04 13:39:56 + cp ${WORKDIR}/${MACHINE}.cmake ${S}/src/machine.cmake Dec 04 13:40:13 that might cause the wrong path to the file as well Dec 04 13:40:23 JaMa: looks like a mistake from me Dec 04 13:40:48 for the cmake the recipe should be alright Dec 04 13:40:58 nyx-modules might need to have some tweaking @ LG end as well Dec 04 13:41:16 Their current implementation is not really ideal for multiple targets like we have Dec 04 13:41:34 Approach suggested by morphis there made more sense to me when I looked at it Dec 04 13:42:38 currently it's a bit of a mix of the two approaches, it's not perfect Dec 04 13:55:58 agreed Dec 04 13:57:52 will need to work a bit more on the qemux86 / virtualbox issues, on Virtualbox 5.2 on ubuntu the image from jenkins doesn't show the UI again, so I've installed 6.0.0_BETA3 there and it doesn't show the syslinux menu + I cannot get the serial port working on ubuntu, so I cannot select it from there -> looks like upgrade VirtualBox to 6 might cause more pain to everybody Dec 04 13:58:10 not sure if they will fix all these issues before 6 is released Dec 04 13:58:49 I haven't tried BETA3 on my Gentoo yet, only BETA2 with its own quirks (like UI not working at all :)) Dec 04 14:12:03 got serial console working and past syslinux menu, still the UI doesn't work Dec 04 14:12:56 modprobe: FATAL: Module vboxvideo not found in directory /lib/modules/4.19.0-yocto-standard Dec 04 14:13:11 looks like I forgot to migrate some changes from my local build to pyro Dec 04 14:14:08 I've triggered unstable qemux86 build just to compare :) Dec 04 17:25:18 Tofe: YGM Dec 04 20:31:48 Herrie: yes, thank you :) Dec 04 20:32:14 I've spotted one or two typos, but the rest looks good to me **** ENDING LOGGING AT Wed Dec 05 03:00:01 2018