**** BEGIN LOGGING AT Sat Jun 15 02:59:57 2019 Jun 15 06:51:14 Morning Jun 15 07:01:06 morning Jun 15 07:30:03 Tofe: For bootd you'd have to fork, add the .service file from https://github.com/webosose/webos-initscripts/blob/master/files/systemd/services/common/bootd.service to the fork, similar to what I did for the others. Or simply add it to the recipe. That's also an option. I'm not a big fan of the whole initscripts repo which seems to be a leftover of the whole legacy init bits. Similar to Jun 15 07:30:03 what I did for rdxd: https://github.com/webOS-ports/rdxd/commit/f159cb54d13777e835015bdbcdf5be46e58fe539 Jun 15 07:31:36 And then also add inherit webos_systemd to the recipe Jun 15 07:32:23 While you're at it you could add PmTrace as well I guess :P ? Jun 15 07:56:36 Tofe: Let me test a few things and PR Jun 15 08:29:37 Tofe: https://github.com/webOS-ports/meta-webos-ports/pull/371 Jun 15 08:32:55 Not run tested, but should do the trick normally Jun 15 09:09:35 Morning! Jun 15 09:09:39 Herrie: oh good, thanks Jun 15 09:11:20 Herrie: looks like the service file depends on /var/systemd/system/env/bootd.env Jun 15 09:22:48 Tofe: Yes, not sure how that one gets installed let me check Jun 15 09:27:32 Tofe: That file doesn't exist in OSE qemux86 image even Jun 15 09:46:10 strange Jun 15 09:53:24 I think it's worth merging and testing Jun 15 09:54:05 ah wait, I can actually at least build-test it Jun 15 09:55:18 Tofe: I build tested it ;) Jun 15 09:55:27 oh ! ok, let's merge then Jun 15 09:55:28 Just no runtime test yet Jun 15 09:55:51 I have a pinephone one, I'll see how it behaves Jun 15 09:57:39 We'd need to see how to deal with these systemd service files Jun 15 09:58:02 It's a bit annoying to having to fork the repos just for this Jun 15 09:58:16 But the initscripts are nasty, so not keen on keeping it either Jun 15 09:58:32 We could supply it as a file instead, which is a bit cleaner, but also not very nice :S Jun 15 09:59:07 I understood why you disabled PmTrace, gives a lot of headaches with kernel I saw? Jun 15 11:59:51 Herrie: the issue with pmtrace was that I wasn't sure how far in the dependencies it would lead us Jun 15 12:00:17 Also it's intricated with performance measuring tools Jun 15 12:00:46 I'm ok to pull the string, but for next release only; otherwise that release will never get out... Jun 15 13:52:05 Tofe: hmmz my VBox image still complains about bootd Jun 15 13:52:54 But I've seen this before it seems it could be due to caching somehow due to same filenames of locally built images, even though I deleted it completely from VBox Jun 15 13:53:07 How is it on PinePhone? Jun 15 13:58:13 Seems I don't have the bootd.service in /lib/systemd/system even thouhg it's in the IPK that I built :S Jun 15 14:06:17 When I manually add the .service I get a bootd.log in /var/log/ Jun 15 14:06:42 Which has the following info: https://bpaste.net/show/b1ab2add3588 Jun 15 14:06:45 Which seems good Jun 15 14:06:55 We didn't include bareapp so that's logical it cannot launch it Jun 15 14:07:17 getForegroundAppInfo didn't exist in our app manager yet probably, so that's also logical Jun 15 14:11:06 We should fix the serialgetty though. It really consumes 98% of my log on qemux86 :S Jun 15 14:11:31 https://bpaste.net/show/bec27c7be229 Jun 15 14:21:47 btw. are you guys on something other than freenode? this thing has been logging me out like crazy lately Jun 15 14:36:20 saidinesh5: Ehm no, not for now ;) Jun 15 14:36:49 You can try another server on freenode? Jun 15 14:36:58 It's been pretty stable here Jun 15 14:50:53 Tofe: Mystery solved: Had another VBox still running :S Jun 15 14:50:55 So it was connecting to that :P Jun 15 15:34:14 Tofe: I think I managed to find a way to solve those permission errors for callback/triggers Jun 15 15:34:19 Testing it now Jun 15 15:34:32 Seems as simple as adding the ls2 calls to the right json file to allow callbacks ;) Jun 15 15:40:12 Herrie: ah interesting. i was using quassel but just found out my freenode was via. irc.ubuntu.com Jun 15 15:40:17 just changed it to irc.freenode.net Jun 15 15:40:30 now fingers crossed for a stable connection Jun 15 16:17:16 Tofe: This should fix those callback/trigger errors in log: https://github.com/webOS-ports/luna-service2-security-conf/pull/2/files Jun 15 16:31:34 bah still getting disconnected i see Jun 15 16:31:45 and interesting.. WARNING: Host distribution "ubuntu-16.04" has not been validated with this version of the build system; you may possibly experience unexpected failures. It is recommended that you use a tested distribution. Jun 15 16:31:56 Herrie: ubuntu 16.04 no longer up for building latest? Jun 15 16:32:40 Herrie: oooh nice !! I have been trying to fix this one for quite some time Jun 15 16:33:32 I don't have write access there, but feel free to merge that Jun 15 16:33:53 For serialgetty, I guess it's a yocto thing Jun 15 16:35:54 https://git.yoctoproject.org/cgit.cgi/poky/plain/meta/recipes-core/systemd/systemd-serialgetty.bb indicated we should have an empty SERIAL_CONSOLES to get rid of it Jun 15 16:41:08 Tofe: poky is what you're using ? or warrior? Jun 15 16:50:26 saidinesh5: We recently moved to warrior and now migrating all the various webOS OSE bits into LuneOS Jun 15 16:50:37 The new LS2 permission model is giving some headaches but getting there Jun 15 16:50:38 Ahh nice Jun 15 16:50:45 LS2 permission model? Jun 15 16:51:17 Tofe: Access should be fixed Jun 15 16:53:31 Ahh interesting.. https://github.com/webosose/luna-service2-security-conf/blob/master/files/sysbus/webos.public.api.json Jun 15 16:59:18 bah stupid freenode Jun 15 16:59:35 ugh Jun 15 17:02:59 saidinesh5: I never had these issues with freenode Jun 15 17:03:24 Tofe: i use a shared quassel server.. so maybe freenode is overloaded with connections from there? Jun 15 17:03:36 saidinesh5: Yeah they basically changes from public/private bus to a single bus model with permissions on API call level, both inbound and outbound, so it's a lot more restrictive and not very well documented at this time, so takes us a bit of time to get our heads around it ;) Jun 15 17:03:56 But from a security perspective it's the right way to do things Jun 15 17:05:06 Herrie: this is for qml/non web apps? Jun 15 17:05:07 yes, the new LS2 security has a finer granularity, and more generic Jun 15 17:05:38 saidinesh5: it's for everyone using the LS2 calls, that being from a webapp or not Jun 15 17:05:54 ohh Jun 15 17:06:34 we can use shortcuts from qml, but the spirit of the thing is that nearly everything goes through LS2 Jun 15 17:07:20 Ahh.. Jun 15 17:07:40 let's see.. compiling again.. iirc the last time i had a system image but some graphics drivers didnt compile or something Jun 15 17:07:46 so all i was stuck with was a shell Jun 15 17:07:54 saidinesh5: which target? Jun 15 17:07:59 intel64-tablet Jun 15 17:08:02 ah yes Jun 15 17:08:08 we have newer mesa Jun 15 17:08:14 nearly 19.1 Jun 15 17:08:15 i have a xiaomi phone to play with too now btw. Jun 15 17:08:17 wow Jun 15 17:08:20 that's quite new Jun 15 17:08:38 yes, that might help with your graphics :) Jun 15 17:08:42 xioami vince Jun 15 17:08:49 which is quite similar to mido i think Jun 15 17:09:12 vince... note 5 pro ? Jun 15 17:09:15 Tofe: would be nice if i can generate a efi bootable build Jun 15 17:09:18 Redmi 5 plus Jun 15 17:09:23 but i think it is teh same as note 5 pro Jun 15 17:09:29 ok, I have a rosy, i.e. note 5 :) Jun 15 17:09:40 oh you got things up and running on note 5? Jun 15 17:09:56 mmmh ah no wait, I'm confused Jun 15 17:10:18 no it's just a redmi 5 Jun 15 17:10:27 damn with their model names Jun 15 17:10:42 Ahh Jun 15 17:10:53 yeah they reuse their model names Jun 15 17:11:00 cancro was Mi 3 and Mi 4 Jun 15 17:11:11 but with minor differences that caused so many headaches for me when maintaining the port back then Jun 15 17:11:34 vince shouldn't be a big challenge, I think Jun 15 17:11:40 mhm it shouldnt Jun 15 17:11:44 I got rosy running within the week Jun 15 17:12:05 (after waiting 2 weeks for unlocking bootloader, that is) Jun 15 17:12:10 but wanted to get the tablet first.. because my Nexus 7 flo is dying and i need a media consumption tablet with enough ram Jun 15 17:12:16 oh of course Jun 15 17:12:22 this one came with unlocked bootloader Jun 15 17:12:27 that's nice! Jun 15 17:12:47 yes, I'm interested in seeing you succeed with the intel tablet too :p Jun 15 17:13:22 yeah.. like last time things were promising for sure.. after creating the franken bootable image Jun 15 17:15:38 yep Jun 15 17:17:56 WARNING: leveldb-1.20+gitAUTOINC+a53934a3ae-r0 do_packagedata: QA Issue: Package version for package leveldb-dbg went backwards which would break package feeds from (0:1.20+git0+a53934a3ae-r0webos1 to 0:1.20+git0+a53934a3ae-r0) [version-going-backwards] Jun 15 17:17:59 interesting Jun 15 17:22:26 Well the Xiaomi devices should be fairly easy Jun 15 17:22:34 Once unlocked Jun 15 17:22:47 yep. unlocked and had a couple of ports already Jun 15 17:23:00 ubports, sailfishos i think Jun 15 17:24:21 saidinesh5: Well I would go down the Halium route, that's easiest. Jun 15 17:24:27 That's what I did for the other targets Jun 15 17:24:29 Ohh Jun 15 17:24:35 Then getting LuneOS bits are easy Jun 15 17:24:42 Herrie: halium route as in ? Jun 15 17:24:46 Just replicate the mido/rosy/tissot bits and adjust were necessary Jun 15 17:24:51 Ahh Jun 15 17:24:59 yeah i still have to replicate things on the lune os side right? Jun 15 17:25:05 https://docs.halium.org/en/latest/porting/first-steps.html Jun 15 17:25:23 Basically add something like: https://github.com/Halium/halium-devices/blob/halium-7.1/manifests/xiaomi_mido.xml Jun 15 17:25:57 Issue seems to be there's not cm14.1 for Vince in TheMuppets though: https://github.com/TheMuppets/proprietary_vendor_xiaomi/tree/cm-14.1 Jun 15 17:27:05 ohh Jun 15 17:27:21 i think with the newer devices the vendor blobs come in their own partition Jun 15 17:27:53 So do we have to worry about TheMuppets repo still? Jun 15 17:28:02 In general yes ;) Jun 15 17:28:06 We still need it it seems Jun 15 17:28:46 yes, we don't depend on any other OS bits now Jun 15 17:29:08 Ahh.. interesting. for cancro the muppets repo mostly provided the files for system/vendor iirc Jun 15 17:29:54 will have to see how they dealt with vince for ubports then Jun 15 17:30:03 Well there are some scripts to extract it from Android I think Jun 15 17:30:14 So when you have Android 7.1.2 image on it you can extract them Jun 15 17:30:19 And upload them somewhere Jun 15 17:30:28 yep Jun 15 17:31:06 thanks for the warning lol. will save me time from having to reflash stock firmware Jun 15 17:32:18 You'll need 7.1.2 for now Jun 15 17:32:35 Newer versions haven't been ported yet. So Android 8/9 bases ROMS won't work Jun 15 17:32:42 Yep.. that thing is still rotting with the stock miui .. no clue which version Jun 15 17:33:02 and yep. iirc the mer middleware was very specific when it comes to android middleware Jun 15 17:33:11 7.1.2 is lineage os 14 right? Jun 15 17:33:20 saidinesh5: vince and whyred should be pretty close, isn't it ? Jun 15 17:33:40 * saidinesh5 checks Jun 15 17:33:48 whyred is note 5 pro I think Jun 15 17:34:33 ah, nevermind, I thought there was already a Muppets for that one, but it's not the case Jun 15 17:35:31 yep Jun 15 17:35:42 whyred is pro.. Jun 15 17:36:38 vince is probably closer to mido Jun 15 17:36:40 Snapdragon 625, Octa-core 2.0 GHz Cortex-A53 Jun 15 17:38:43 Yeah pretty close then Jun 15 17:38:45 Tofe: https://github.com/webOS-ports/meta-webos-ports/pull/372 Jun 15 17:38:55 That should solve some more LS2 related errors in logs hopefully Jun 15 17:39:26 will i have to put up with these LS2 errors too when booting warrior? Jun 15 17:39:59 Tofe: Ah let me add another one ;) Jun 15 17:40:03 For bootd Jun 15 17:41:11 bootd is their systemd replacement? Jun 15 17:41:51 Tofe: Nevermind that one doesn't seem critical Jun 15 17:41:55 I'll leave it for now Jun 15 17:42:04 saidinesh: Well no, it's their systemd helper ;) Jun 15 17:42:24 ohh Jun 15 17:42:27 systemd helper? Jun 15 17:42:28 They have some nasty hybrid between systemd and previous system Jun 15 17:42:45 ohh Jun 15 17:42:47 We try to stay away from it as far as we can, but we cannot avoid it completely it seems Jun 15 17:43:09 as long as it brings us plain text logs... Jun 15 17:43:20 does it? Jun 15 17:43:34 In general those LS2 errors won't bug you too much, they appear in the logs and have some impact on things wokring, but stuff boots and we get UI and most things work Jun 15 17:43:37 So nothing too shocking Jun 15 17:44:28 Just some more advanced stuff like adding a Google C+DAV account, syncing Contacts, Calendars etc is impacted ;) Jun 15 17:44:29 Ahh as long as i will see a GUI this time Jun 15 17:44:42 But that's already far into the OS ;) Jun 15 17:45:22 Audio, bt, sensors, browsing, wifi etc work ;) Jun 15 17:45:51 oh wifi + browsing + audio + video is all i need Jun 15 17:45:55 from this tablet Jun 15 17:49:08 The whole OSE migration is 5 steps forwards, 2 backwards Jun 15 17:49:18 backwards as in? Jun 15 17:49:41 Well stuff that temporary broke for now ;) Jun 15 17:49:55 Ahh that's okay.. as long as it is a "cleanup" overall Jun 15 17:50:48 Well it's 5+ years of internal LG development on components + real life testing @ LG on millions of TV's, so in terms of overall stability it should be an improvement ;) Jun 15 17:51:59 bah Jun 15 17:52:03 why freenode why Jun 15 17:57:46 I don't have such issues really Jun 15 17:57:56 Got ZNC running on my builder and it's pretty stable Jun 15 17:58:01 Hardly any disconnects Jun 15 18:05:30 * saidinesh5 shrugs Jun 15 19:09:45 Tofe: Somehow bootd.service doesn't seem to start for me Jun 15 19:17:19 Though file seems to be there in right location Jun 15 19:59:26 bah humbug: ERROR: qtwayland-5.12.3+gitAUTOINC+bcdc00bcdc-r0 do_configure: QA Issue: qtwayland: invalid PACKAGECONFIG: drm-egl-server [invalid-packageconfig] Jun 15 20:21:29 Tofe: Seems I managed to kill all the "ls-hubd LSHUB_NO_OUT_PERMS {"DEST_APP_ID":"com.webos.lunasend-532","SRC_APP_ID":"com.palm.configurator","EXE":"/usr/sbin/configurator","PID":506} "com.palm.configurator" does not have sufficient outbound permissions to communicate with "com.webos.lunasend-532" (cmdline: /usr/sbin/configurator -c {"log":{"appender":{"type":"syslog"},"levels":{"confi Jun 15 20:21:29 gurator":"notice"}}} service)" locally :D Jun 15 20:21:37 Let me prepare a PR and call it a day :P Jun 15 20:24:18 OK I'll leave that to tomorrow. Seems my patch for db8 would need to be merged anyway ;) Jun 15 20:31:16 Herrie: thanks ! here are, also, some notes of mine: https://bpaste.net/show/13ddc476bfd0 Jun 15 20:34:32 Tofe: Seems that's what I did Jun 15 20:34:50 ok :) Jun 15 20:35:07 I just noted in a file the local modifications I did on device last week Jun 15 20:35:26 Let's see if we can sync Google stuff now Jun 15 20:35:27 I never had time to send it "upstream" Jun 15 20:35:40 That's always a good test Jun 15 20:35:44 yes :) Jun 15 20:35:57 I.e. Contacts & Calendar Jun 15 20:36:27 If the Calendar App icon shows date that's a good sign too Jun 15 20:36:45 For the scrolling of accounts, I'm really not sure where to look for, I must say Jun 15 20:38:00 It's probably in the Enyo 1 is file somewhere Jun 15 20:38:11 I recall we did some fixes there previously Jun 15 20:38:21 s/is/js Jun 15 20:41:19 ok Jun 15 20:46:27 Let me see if I can find a link tomorrow Jun 15 20:47:54 yeah it's bed time for me too :) Jun 15 20:48:06 https://github.com/webOS-ports/enyo-1.0/commits/webOS-ports/master Jun 15 20:48:30 Some of the commits from morphis might be worth to check Jun 15 20:49:33 I.e. https://github.com/webOS-ports/enyo-1.0/commit/99a58156d16ca7f0a1b1e4214aeba04c07570b38 Jun 15 20:50:07 And https://github.com/webOS-ports/enyo-1.0/commit/26b98a0179ffb4f6a8bbf1b72a59bcda8e5a654f Jun 15 20:56:55 yes, right **** ENDING LOGGING AT Sun Jun 16 02:59:58 2019