**** BEGIN LOGGING AT Mon Jun 24 02:59:57 2019 Jun 24 06:42:45 Morning! Jun 24 07:25:51 Morning! Jun 24 07:47:23 Ooh some nice Monday updates from OSE ;) https://github.com/webosose/meta-webosose/commits/master Jun 24 07:47:48 Seems to be Sumo ;) Jun 24 08:32:06 hehe Jun 24 10:15:15 Seems there's a Thud branch now too ;) Jun 24 11:03:10 ah, they are so lagging behind ;) Jun 24 11:29:37 Well Thud is a lot better v.s. Morty :P Jun 24 12:07:06 Hmmz RaspberryPi4 has interesting specs :) Jun 24 12:13:59 yes, it's quite powerful; but, it's not as power efficient as before ;) https://www.raspberrypi.org/magpi/wp-content/uploads/2019/06/2019-06-21-15_44_14-Window.png Jun 24 12:19:05 True but compared to 3B not a big increase Jun 24 12:19:11 3B+ I mean Jun 24 12:26:38 Do you know, my ZNC is running on a RPi A machine :p Jun 24 12:34:21 Hehe ;) Jun 24 13:03:42 morning Jun 24 13:06:00 JaMa: Morning! Jun 24 13:06:05 Seems things are hitting public GH ;) Jun 24 13:29:45 yes, finally :) Jun 24 13:30:51 I merged a bunch of older OSE migration bits in the weekend. Have a few questions for other things to migrate as well where you might be able to give some guidance Jun 24 13:31:17 Just seems I caused an issue with the signatures again :S Jun 24 13:31:24 Hash for dependent task luna-downloadmgr/luna-downloadmgr.bb.do_package changed from 5967feec828a5ec0202605af9e8c6bda42573c41278e8c5dfa62cae540daef60 to 3f64fcee94a38b5689b9381b0d13a477e1b3c5f2e57411e6ac583d812e6d60d8 Jun 24 13:33:17 OSE has https://github.com/webosose/meta-webosose/blob/master/meta-webos/recipes-webos/luna-downloadmgr/luna-downloadmgr.bb#L20 Jun 24 13:33:26 Which I dropped in our recipe. Is that the cause? Jun 24 13:34:10 did you drop it druing the build? then definitely yes Jun 24 13:34:21 Not during build ;) Jun 24 13:34:50 different PACKAGE_ARCH causes do_package to change signature Jun 24 13:35:05 where are you seeing the message about changed hash? Jun 24 13:36:39 JaMa: http://jenkins.nas-admin.org/job/LuneOS/job/luneos-testing_workspace-compare-signatures/136/console Jun 24 13:38:05 The other ones (libblockdev, ndctl, oprofile and umip) where there before already it seems Jun 24 13:39:41 yes, libblockdev, ndctl, oprofile and umip are waiting for next meta-oe/warrior update Jun 24 13:40:28 and yes it was wrong to drop the inherit, see following lines https://github.com/webosose/meta-webosose/blob/master/meta-webos/recipes-webos/luna-downloadmgr/luna-downloadmgr.bb#L23 and L24 Jun 24 13:40:48 I wasn't 100% sure what those did ;) Jun 24 13:40:53 Just previously we mainly dropped them Jun 24 13:40:57 I'll add it back then Jun 24 13:40:58 if you don't remove the dependency on MACHINE variable from the component, then dropping the inherit only makes it worse Jun 24 13:41:50 you should check in the component respository why they use MACHINE variable Jun 24 13:42:04 I don't think it should do that really Jun 24 13:42:11 and if possible, then rework that Jun 24 13:42:43 in OSE (and OWO) everything is MACHINE_ARCH by default, so they inherit webos_machine_dep without much thought Jun 24 13:42:57 Ah seems that there's an optional MACHINE specific config file.... https://github.com/webosose/luna-downloadmgr/blob/master/CMakeLists.txt#L147 Jun 24 13:43:13 in LuneOS we're better and we build most of stuff as TUNE_PKGARCH so we need to be careful and fix the design flaws they don't care about Jun 24 13:43:40 So we should put that config file in a separate repo/recipe like we did with others and make that recipe MACHINE_ARCH Jun 24 13:43:50 https://github.com/webosose/luna-downloadmgr/tree/master/files/conf Jun 24 13:44:04 Such as https://github.com/webOS-ports/meta-webos-ports/blob/warrior/meta-luneos/recipes-webos/luna-sysmgr-conf/luna-sysmgr-conf.bb ? Jun 24 13:44:09 if we really need different config as you can see it has config for MACHINEs not supported for very long time Jun 24 13:44:55 I don't think we do really ;) Jun 24 13:45:02 So we could just drop 23 and 24 here Jun 24 13:45:09 yes Jun 24 13:45:23 these configs only set different WANInterfaceName Jun 24 13:45:46 but all new MACHINEs are using the default config already, so it probably isn't really needed Jun 24 13:45:51 Yup Jun 24 13:45:56 OK will drop it then from recipe Jun 24 13:46:06 BeAggressive__Bee_Eee_AGGRESSIVE=true Jun 24 13:46:12 looks a bit strange as well :) Jun 24 13:47:11 Probably someone at Palm/HP being funny: https://github.com/webosose/luna-downloadmgr/blob/47c7c3cc9e425fed4da71e62e7f6e6c08a197f7f/src/DownloadSettings.cpp#L230 Jun 24 13:47:16 Seems valid though :P Jun 24 13:52:40 JaMa: You know something about ecryptfs? When I try to add appinstalld2 which DEPENDS on ecryptfs I get the following in my logs: https://bpaste.net/show/163f32f02e84 Jun 24 13:52:57 I guess we'd need to add the defconfig changes to our qemux86 & other kernels? Jun 24 13:53:18 Similar to https://github.com/webosose/meta-webosose/blob/master/meta-webos/recipes-kernel/linux/linux-yocto/crypto.cfg Jun 24 13:55:45 these weren't introduced because of ecryptfs, but might be related Jun 24 13:56:37 Ah OK Jun 24 13:56:59 and from ecryptfs, OSE is using only the ecryptfs-utils, not sure if ecryptfsd should actually be installed in the image Jun 24 13:58:01 btw: 9c9687d34 Anbox: Add elfutils to DEPENDS to solve QA issue Jun 24 13:58:10 have you seen this QA issue in your builds somewhere? Jun 24 13:58:19 I've seen it only with master/unstable builds Jun 24 13:58:21 JaMa: Yes in my local build Jun 24 13:58:26 and similar commit is already in master branch Jun 24 13:58:33 Where I added some OSE bits Jun 24 13:58:44 ok, strange, not sure what causes it Jun 24 13:58:55 More than what was merged. I left all the "controversial" changes out of my PR ;) Jun 24 13:59:02 So only straight forward ones went in Jun 24 13:59:08 Other ones we'll do after release ;) Jun 24 13:59:14 unless something in anbox dependency tree newly gained dependency on elfutils as well Jun 24 13:59:22 Not to introduce new regressions Jun 24 13:59:28 and then anbox might have started autodetecting it from RSS Jun 24 14:01:25 did you forget to remove ./meta-luneos/recipes-luneos/services/downloadmanager.bb ? it doesn't seem used anymore Jun 24 14:03:36 Not used anymore yes, can delete it, will do that ;) Jun 24 14:04:34 the change in meta-luneos/recipes-webos/novacomd/novacomd.bb seems to reintroduced upstart init script, I don't think we want that Jun 24 14:06:47 JaMa: Ah I didn't realise this was an upstart script, yes let's get rid of that too :P Jun 24 14:08:58 I can drop the FILESEXTRAPATHS_prepend as well then right? Jun 24 14:09:31 yes Jun 24 14:09:51 FILESEXTRAPATHS_prepend was wrong with or without the upstart script anyway Jun 24 14:11:19 Ok both fixed Jun 24 14:12:18 JaMa: Any easy way to see the final defconfig that's being used in the qemu kernel? Jun 24 15:10:10 Herrie: adding the .cfg files? Jun 24 15:10:46 see https://github.com/webosose/meta-webosose/commit/9d3c9ceb9ed076cd4f4ecb9ab15f9cfc9ca7534b if that's the same issue as what you're seeing Jun 24 15:35:14 JaMa: Yes partially, but I also would like to see if there's anything else in qemu's defconfig, so would like to know the whole defconfig it's using Jun 24 15:36:26 most reliable one is the .config file in work-shared Jun 24 15:37:02 tmp-glibc/work-shared/qemux86/kernel-build-artifacts/.config Jun 24 15:41:51 OK :) Jun 24 15:41:57 That's what i was looking for :) Jun 24 16:42:33 Tofe: Pushing some small fixes for things I see in the logs Jun 24 16:42:46 Doesn't solve the email or contacts problem yet, but doesn't hurt to have them addressed either Jun 24 17:02:52 Updated the filecache one a bit Jun 24 17:03:12 Seems it sometimes gets called from LunaWebappMgr with null Jun 24 17:03:16 So added "" Jun 24 17:03:26 We had that previously as well in some places Jun 24 17:06:48 Proper fix should be to figure out why there's a null I guess Jun 24 17:27:57 Herrie: some of our code registers on the LS bus with an empty string identifier Jun 24 17:29:07 https://github.com/webOS-ports/luna-qml-launcher/blob/master/src/lunaqmlapplication.cpp#L155 for instance Jun 24 17:29:35 and I guess somewhere in LunaWebAppMgr we do the same Jun 24 17:30:45 https://github.com/webOS-ports/luna-webappmanager/blob/c4f72bab46d9785af147cf1d4f6af0df0e4bab66/src/activity.cpp#L65 might be one Jun 24 17:31:24 but I'm still not very at ease with the whole LS2 APIs Jun 24 17:53:52 Tofe: Yeah Same here Jun 24 17:57:54 LS2 still bit mystery Jun 24 17:58:33 Not sure why we send empty identifier? Jun 24 18:10:17 morphis probably had his reasoning **** ENDING LOGGING AT Tue Jun 25 03:00:38 2019