**** BEGIN LOGGING AT Sun May 24 02:59:59 2020 May 24 03:04:13 hi. can i pause build process? Facing power cuts. May 24 03:06:05 on ups now. battery will last only some time May 24 04:49:15 BT40: Hi May 24 04:49:28 No real way to pause it except for CTRL+C and restart later May 24 04:49:45 It will then only build the components that were not build yet May 24 04:49:56 1 time CTRL+C it will finish what it's still building May 24 04:50:01 And not start new tasks May 24 04:50:25 2 times CTRL+C will abort immediately (not really recommended since it can cause some artifacts and small issues when restarting build). May 24 04:50:36 Fixable usually, but for a beginner it can be a pain May 24 04:53:17 In terms of features that work, not sure you saw my message: Wifi, BlueTooth, Audio, sensors work (rotation, gyro, light sensors, leds etc work). No hardware GPS support at the moment only location via WiFi (Google/Mozilla), though it shouldn't be rocket science to add it going forward. Browser is basic but based on pretty recent Chrome (80 in latest testing build!). SMS, calls, data May 24 04:53:17 mostly works, depends a bit on device. Calculator, notes work. May 24 04:54:10 Stuff like Skype, Telegram is hit & miss a bit for the moment. Backend is there, UI in the app needs some TLC, but there is an option to sideload the legacy variant (which we cannot distribute because it was never open sourced by HP/LG). May 24 07:38:18 Morning! May 24 07:40:41 As for calls, we still have issues on Halium 7.1 devices at the moment; I think it's just an audio routing issue, but still the sound isn't there during calls. May 24 08:35:09 Tofe: Going through the various repos in my ACG branch and updating them one by one May 24 08:35:15 Quite tedious, but getting there May 24 08:35:52 Herrie: I didn't do any merge yet, not knowing exactly which ones were already ready. But the changes so far seems fine to me. May 24 08:40:49 There will be plenty more, some tidying up, it might actually solve the remaining issues in the logs and maybe also the wallpaper issues, let's see... Hope to finish it today or tomorrow May 24 08:50:51 Hi. My initial porting was interrupted due o power cut. I resumed the same , increased the threads from 4 to 12 in local.conf. it proceeded faster than before. everything was fine. Then again interrupted by power cut May 24 08:51:10 Second time i resume, the threads automatically got reduced to 4 May 24 08:51:36 when i checked local.conf, it was changed and thread definition was commented. May 24 08:52:13 I again changed that file, setup environment again. But still i can get 4 threads max. Can someone guide May 24 08:52:45 something is changing local.conf file automatically. My changes to threads is being overriden May 24 08:53:19 These are the lines: PARALLEL_MAKE = "-j 4" May 24 08:53:19 BB_NUMBER_THREADS = "12" May 24 08:53:45 earlier this worked fine, after secont power cut, now not working May 24 08:56:03 # PRSERV_HOST = "localhost:0" This is also bein automatically commented in local.conf. I uncommented earlier on suggestion of you guys. Want to understant what thing is rebuilding local.conf? May 24 09:04:03 BT40: make sure to change the correct local.Conf May 24 09:04:07 There are 2 May 24 09:08:35 BT40: I'd say 4 threads and 4 parallel make is the maximum reasonable value for your system. I've these ones, and my CPU is already maxed out May 24 09:10:16 Herrie: yes, the wallpaper thing is probably some getPreferences call which gets refused May 24 09:11:52 i changed both one. When build start, i come back and see, conf is reverted automatically. something is altering my changes behind the scenes Herrie May 24 09:12:34 Tofe, I also assumed that 4 threads are max theoretically. But with 12, build time is around 3 times faster in my case May 24 09:13:50 Because many times, 4 threads all are devoted to downloading and cpu sits idle. By 12, some of them are allocated to compiling and performance increases in my case May 24 09:16:14 ok then :) May 24 09:17:42 also while doownloading, sometimes 4 threads dont utilize my whole bandwidth due to server side limitations. 12 threads make sure my more network is utilized. this also improves build time May 24 09:20:03 as long as your HDD isn't a bottleneck and it speeds up things, that's fine May 24 09:20:46 just beware when the downloads will be finished: you could end up with 12x4 instances of gcc at a given time, and that could stress your available RAM a bit May 24 09:21:31 so wandering what is changing local.conf May 24 09:21:50 earlier that worked fine May 24 09:22:06 I quite don't see, there isn't any automagic update of that file with bitbake May 24 09:22:26 BT40: When you restart build you just need to do ". setup-env" and then MACHINE=land etc May 24 09:22:29 you're on the right conf/local.conf file ? May 24 09:24:00 Herrie, yes, after that conf file is altered by system automatically. May 24 09:25:11 yes tofe, i implement changes on both files. After running two commands . setup-env, file is altered May 24 09:27:15 Did you save it? Maybe you need sudo to save it? May 24 09:27:22 Happens in some folders May 24 09:28:23 ok i will reconfirm May 24 09:31:22 checked, sudo is not the issue. Changes are saved when i save. Files are altered after . setup-env and make update command May 24 09:47:48 Don't DO make update! May 24 09:47:52 That updates it May 24 09:47:57 Just . setup-env May 24 09:48:09 Then right away MACHINE=land bb luneos-dev-package! May 24 09:48:28 okay May 24 09:52:49 tested this way also, still 4 threads. although contents of file are not changed. May 24 09:54:42 make update command is necessary too implement changes it seems, otherwise it assumes previous settings May 24 10:04:17 Not necessary May 24 10:04:31 Make update is to pull updates from remote repos May 24 10:04:39 Not necessary for now May 24 10:05:28 I hardly ever use make update while developing May 24 10:05:47 ok. so what is holding threads? Earlier same method worked. Now not working. May 24 10:06:37 Strange thing is, first time also i run make update command since i didn't knew what its function is. It didn't made changes that time. May 24 10:07:04 same commands, different results. :) May 24 10:07:44 surely there might be some mistake i am making. Or maybe due to abrupt interupt due to sudden computer off. maybe......who knows May 24 10:10:52 There is a local.conf file in temp-glibc folder. I will try to delete that and see if that might be causing issue. May 24 10:12:32 deleted, still no improvement May 24 11:05:14 BT40, you can try to set 2 threads just to make sure the changes count. There is a chance that only 4 packages can be built at this time, and all others depend on them (waiting).. just thought. May 24 12:47:29 Tofe: I think I might have found why emails don't sync right away and only after reboot :) May 24 12:47:40 Seems I probably forgot to set some ACG permissions somewhere May 24 12:47:52 In mojomail, so will test that as well. May 24 13:51:06 . setup-env May 24 13:51:06 environment: TOPDIR not defined May 24 15:18:29 Bt40 is does ring a faint bell May 24 15:18:37 Did you try a system reboot? May 24 15:19:48 Tofe: seems that at some point we need to merge some parts of powerd into sleep. Seeing this is all c(++) code I'll leave it to you or nizovn at some point. May 24 15:20:04 Shouldn't be rocket science probably but don't know May 24 15:38:36 Tofe: Now started updating the recipes, I guess I can start a build soon :) May 24 16:08:36 Herrie: I never really looked into sleepd or powerd, but it didn't look very complex May 24 16:09:07 Tofe: No it isn't.. Seems most from OWO powerd is now in sleepd, but not all it seems May 24 16:09:29 Some battery & charger stuff not it seems, but seems it should be fairly easy to lift & drop May 24 16:09:53 bt40: TOPDIR is supposed to be defined in conf/topdir.conf May 24 16:09:57 I see references to a service called com.webos.service.battery or something. So could be they placed it there but didn't open source it (yet) May 24 16:10:43 Herrie: they may have had no use for it so far, TV rarely have batteries May 24 16:11:55 when you'll have something that builds, I'll try it too on tissot May 24 16:11:58 Tofe: True, but their robots shoudl have ;) May 24 16:12:19 TOfe: Well I was toying with your fancy qml-settings files and it didn't like my changes... I'll switch back to old stuff for now May 24 16:12:26 It's not that critical for the moment May 24 16:12:53 ah, yes, maybe I went a bit too far with qml settings apps, it's not very easy to tinker with May 24 16:14:38 Well it's the cmake and .in.in files May 24 16:14:44 Just need to revisit May 24 16:14:56 Nothing critical for now really May 24 16:15:11 For any of the issues we were seeing it's not critical to include May 24 16:19:51 Ok that seems better May 24 16:22:08 Making rootfs now May 24 16:22:12 So can test soon May 24 16:27:04 And I broke it LOL May 24 16:27:40 Already see a type in json May 24 16:27:48 should've run everything through jsonlint.com LOL May 24 16:27:58 I did for some but not for all May 24 16:31:53 :) May 24 16:43:50 But should be easy enough to fix May 24 16:43:59 But no guarantee it will boot then of course May 24 17:30:15 And another syntax error... Leading , before ], no good ;) May 24 17:30:26 Fixed those, no back to not having lists in FirstUse May 24 17:30:29 Will check logs May 24 17:41:31 Tofe: Ah seems I forgot to put the QML container ;) May 24 17:42:38 WebApps launch, so that's good ;) May 24 17:42:44 Still no background though May 24 18:28:12 Herrie: so it's about the same as before May 24 18:29:08 Tofe: I would need to compare side by side, but likely yet May 24 18:29:10 Yes May 24 18:29:20 maybe I can try to debug this background thing on my current tissot, it'll probably be the same fix on your newer build May 24 18:29:38 Just need to fix luna-qml-launcher May 24 18:29:56 Tofe: Well I fixed quite some permissions of things May 24 18:30:23 Hard to say what it will impact in the end May 24 18:30:34 But might be the same fix May 24 18:30:38 I'll look at it anyway, a rebuild would mean 5h of build, so no analysis of the issue until tomorrow May 24 18:31:07 Seems we also need to pass some params to qt for loading local files going forward it seems May 24 18:31:39 We have some /usr/palm check in qml somehere May 24 18:35:54 I mean some qrc or whatever prefix they use pointing to /usr/palm which causes issues May 24 18:36:06 Maybe not in 5.15 but in future May 24 18:36:17 Saw this mentioned in logs somewhere May 24 18:36:44 Don't remember journal or /var/log/messages May 24 18:41:36 For wallpaper, I think the hint is "DEBUG: 04:51:46.745: Failed to call preferences service: Failed to call remote service: sh != NULL" from luna-next May 24 18:42:03 meaning there is an issue with the service in LunaSysAPI/Preferences.qml May 24 18:43:06 Tofe: But I think this was there already for a long time May 24 18:43:19 It sounds very familiar to me at least May 24 18:43:19 oh? May 24 18:43:38 ok May 24 18:43:56 I can check a bit later, but it could be a lead nevertheless May 24 19:45:44 mmmh we send the getPreferences call to LS2, LS2 answers (I can see the good answer with ls-monitor), but in C++ we aren't called on our callback May 24 19:58:11 as weird as it looks, I can't exclude a bug in luna-service2 library... May 24 20:18:23 Herrie: ok, I fixed it, it was just because our LS2 QML adapter was still using LS2RegisterPubPriv May 24 20:18:57 Ah there are more of these May 24 20:19:10 We should grep and replace them May 24 20:19:20 Great it's fixed! May 24 20:20:34 Herrie: do you already have an ACG branch on luneos-components I could push to ? May 24 20:20:44 I see a qt-5.15 branch May 24 20:27:57 Herrie: pushed to tofe/acg : https://github.com/webOS-ports/luneos-components/commit/33d6ec3a66d02efeb354700c07cc6a5751e0a9e8 May 24 20:41:13 Great thnx **** ENDING LOGGING AT Mon May 25 02:59:57 2020