**** BEGIN LOGGING AT Tue Mar 20 03:00:07 2018 Mar 20 06:20:01 Morning! Mar 20 07:55:01 Morning! Mar 20 07:55:10 Herrie: do you have the list of users somewhere? Mar 20 07:55:36 we should just check that there isn't any conflict with the usual Linux users/groups Mar 20 08:28:23 Tofe: I put something together let me find it Mar 20 08:28:32 It doesn't conflict since it's above the 1000 Mar 20 08:28:50 *nix uses up to 999 for the users Mar 20 08:29:05 base_passwd even changes some standard one to match the android ones Mar 20 08:29:23 I.e.: https://github.com/shr-distribution/meta-smartphone/blob/pyro/meta-android/recipes-core/base-passwd/base-passwd/change-android-group-ids.patch Mar 20 08:38:25 Tofe: https://docs.google.com/spreadsheets/d/11bs83StV1pJbW5AXDT7a8Im0wsD-MeJObzn4yFDLSPU/edit?usp=sharing Mar 20 08:38:30 This is what I compiled from 5.1 + 7.1 Mar 20 08:38:53 There are some manufacturers like Motorola which seem to add soem additional users of their own Mar 20 08:39:27 These are the common ones defined in android_filesystem_config.h for Halium 5.1/7.1 Mar 20 08:41:24 I suspect missing 3011 (sensors) might be the cause of my issues with sensors for example Mar 20 08:41:40 Ah that one isn't on my list even :P Mar 20 08:43:35 #define AID_SENSORS 3011 /* access to /dev/socket/sensor_ctl_socket & QCCI/QCSI */ Mar 20 08:44:23 With manufacturer specific id's I mean: https://github.com/LineageOS/android_device_motorola_clark/blob/cm-14.1/include/private/android_filesystem_config.h#L119 Mar 20 08:44:35 I guess we could use a bbappend for those as well Mar 20 08:44:44 In the specific machine Mar 20 08:47:14 Ah the order on Halium is a bit funny Mar 20 08:47:15 https://github.com/Halium/android_system_core/blob/halium-7.1/include/private/android_filesystem_config.h#L132 Mar 20 08:47:21 Should've been a few lines up ;) Mar 20 11:44:24 Herrie: is there a way to get them automatically ? during Halium build, or something similar ? Mar 20 11:47:06 Tofe: That's a good question... I guess Halium will always provide the standard android_system_core it seems. So the vendor specific id's would be missed there, not sure that would cause issues.... Mar 20 11:47:29 not sure either Mar 20 11:49:56 Tofe: But in general these id's don't change a lot Mar 20 11:50:11 So we should be OK with adding the ones from android_system_core Mar 20 11:50:29 And if anything specific is required it could always be done with the device specific bbappend Mar 20 11:51:05 ok Mar 20 12:12:31 JaMa: Hi Mar 20 12:12:38 If I simply remove avoutput-adaptation-layer-mock I get other errors Mar 20 12:12:42 What do I need to do exactly? Mar 20 12:13:09 what kind of errors? Mar 20 12:13:13 I worked around the Node-gyp issue by actually installing the NodeJS version on my host machine Mar 20 12:13:20 about missing avoutput-adaptation-layer-mock? Mar 20 12:13:25 JaMa: Yes Mar 20 12:13:31 JaMa: Chromium is now just building Mar 20 12:13:36 you will need to remove whole avoutputd Mar 20 12:13:49 for qemux86 Mar 20 12:15:27 JaMa: OK. umediaserver doesn't use it? Mar 20 12:51:27 it should work without it on qemu86, or you can just set empty VIRTUAL-RUNTIME_aval-impl Mar 20 12:51:34 it's only runtime dependency from avoutputd Mar 20 12:57:41 JaMa: OK will try when my RPi3 image finishes building Mar 20 13:05:17 JaMa: We're looking into which components make sense to update to webosose version. I guess eventually it'll be all of them, but the lack of commit history, code styling & copyright changes makes it quite tedious. Any suggestions from your side? Mar 20 13:05:39 Except for pulling everything in as a separate branch, comparing them with ?w=1 to ignore whitespaces? Mar 20 13:06:07 Everything in webosose was pushed in a single commit Mar 20 13:06:50 Herrie|Laptop: There might be some references to internal LG stuff in the commit history, and I guess they didn't want to risk exposing it Mar 20 13:07:10 Tofe: Likely Mar 20 13:07:44 Nevertheless, it's going to be tough using this code Mar 20 13:11:59 Tofe: This one has a lot of changes, but seems fairly low impact: https://github.com/webOS-ports/nyx-lib/compare/webOS-ports/master...webOS-ports:webosose?expand=1&w=1 Mar 20 13:13:49 Same with https://github.com/webOS-ports/nyx-modules/compare/webOS-ports/master...webOS-ports:webosose?expand=1&w=1 Mar 20 13:14:00 Lots of new files, but should have little impact on existing bits Mar 20 13:14:05 Not too much code changes Mar 20 13:15:06 yes, hopefully some of the repos are fairly easy to analyze Mar 20 13:17:02 Same goes for https://github.com/webOS-ports/nyx-utils/compare/master...webOS-ports:webosose?expand=1&w=1 Mar 20 13:17:27 I think based on quick analysis this could be a good candidate to start with for example. Or db8 Mar 20 13:17:36 The DB8 changes normally would have little impact on the system Mar 20 13:21:26 yes, I was complaining about the same (not including the history), but it was being updated for long time without any expectation to open it again and the history is full of internal references and proprietary code which was removed only before the webOS OSE release Mar 20 13:21:55 JaMa: OK... Just changing the 30 components we use isn't going to be easy now ;) Mar 20 13:22:06 We're still figuring out what's the best approach... Mar 20 13:22:11 Any suggestions are welcome Mar 20 13:22:33 but LuneOS is probably the only project which really cares about comparing with old stuff Mar 20 13:23:21 have you seen the webOS OSE news on engaget? they mentioned that LuneOS is replacement for Open webOS :) Mar 20 13:23:45 https://www.engadget.com/2018/03/19/lg-webos-open-source-edition/ Mar 20 13:23:56 (It's not to be confused with another open-source version of webOS, OpenWebOS, which became LuneOS for mobile devices in 2014.) Mar 20 13:25:35 JaMa: Yeah LOL Mar 20 13:25:38 Saw that Mar 20 13:26:11 Tofe: Stuff like DB8 might have a lot of code changes, but that's mainly internal DB code & optimizations, the interaction with it should have little change Mar 20 13:27:21 I don't have any specific suggestions, IMHO not worth starting to rebase all our forks on top of what was release now, I would start with components you're not happy with in LuneOS and checking if it was fixed/improved in webosose side Mar 20 13:33:21 JaMa: OK Mar 20 13:49:04 Is there some good video with LG's webOS in action ? Mar 20 13:49:33 I wonder how the apps look like, how the compositor handles windows and notifications, that kind of stuff Mar 20 13:53:32 Tofe: Nothing yet I guess Mar 20 13:54:55 My Chromium should be done shortly, then I can try the qemux86 I gues Mar 20 13:55:02 ok :) Mar 20 13:55:44 I mean to try to build qemux86 Mar 20 13:55:55 Before I have an image and get qemu itself running will be a while Mar 20 13:56:28 I noticed interesting URL in Youtube on the webosose: https://www.youtube.com/tv/ Mar 20 13:56:53 Seems to be a layout that could work fine on a tablet Mar 20 13:56:56 Not sure on phone Mar 20 13:57:02 interesting Mar 20 13:59:15 Found it here: https://github.com/webosose/test-apps/blob/master/com.webos.app.test.youtube/index.html Mar 20 14:22:40 Tofe: I'm going through the interesting repos 1 by 1. I assume you have no issues that I already migrate the ones that have no code changes at all (except for license) ? Mar 20 14:25:19 Seems that novacomd for example didn't change Mar 20 14:26:45 I'll of course PR to review Mar 20 14:40:11 JaMa: Do we need to replicate all the new classes when we want to use them? Mar 20 14:40:48 I.e.: https://github.com/webosose/meta-webosose/tree/master/meta-webos/classes or there's a more clever way to achieve this? Mar 20 14:42:32 Herrie|Laptop: of course, please do Mar 20 15:02:43 Tofe: Only 1 screenshot so far: http://forum.webosose.org/t/webos-raspberry-iso/192/3 Mar 20 15:05:28 Herrie|Laptop: I could also run the gallery partially Mar 20 15:06:17 Ok Chromium & image finished for RPi3, now building qemux86 Mar 20 15:07:51 Gallery? Mar 20 15:08:30 https://github.com/webosose/qml-webos-framework/tree/master/examples/gallery Mar 20 15:09:07 but it required quite some tweaking to run it on a windows desktop :p Mar 20 15:16:23 Not surprised :P Mar 20 15:41:07 Herrie|Laptop: only the new bbclasses if you want to use some new components inheritting them e.g. enactjs.. Mar 20 15:41:29 Herrie|Laptop: the rest will have only a few changes and IIRC nothing really needed for LuneOS Mar 20 15:42:56 JaMa: OK Mar 20 15:43:01 I'll cherry pick then Mar 20 15:53:57 JaMa: Somehow didn't get that qemu to work Mar 20 15:54:22 I removed the whole avoutput folder now and removed the packageconfig for it from umediaserver Mar 20 15:54:28 Now it seems to continue building at least Mar 20 16:00:27 have you tried setting empty VIRTUAL-RUNTIME_aval-impl as suggested before? that should be the only dependency Mar 20 16:02:33 It seems to reset it..... Mar 20 16:02:46 I.e. I change and save, then I run make webos-image and it resets it Mar 20 16:02:56 I'm not very familiar with LG's implementation here :P Mar 20 16:03:01 I liked the plain simple bb more :P Mar 20 16:04:58 eh? where do you set it? Mar 20 16:05:16 it's still plain simple bb Mar 20 16:05:47 https://github.com/webosose/meta-webosose/blob/master/meta-webos/recipes-webos/avoutputd/avoutputd.bb#L9 Mar 20 16:05:51 only mcf can modify the recipes Mar 20 16:05:59 especially if you use --clean Mar 20 16:06:02 Ah ok Mar 20 16:06:08 Maybe I mcf-ed in between Mar 20 16:06:49 then you got mcf-ed by mcf :) Mar 20 16:07:22 Anyway it's @ 54% now, so it'll still be a while Mar 20 16:07:36 Qemu is installed my laptop, so when I'm home I can play and investigate :P Mar 20 16:07:41 now running with umediaserver PACKAGECONFIG removed? Mar 20 16:07:56 keep in mind that returning that will rebuild chromium again Mar 20 16:08:26 while changing avoutputd RDEPENDS shouldn't Mar 20 16:09:42 Well Chromium didn't start yet Mar 20 16:09:46 Let me try something else then :P Mar 20 16:09:47 One of the interesting parts we can take from webosose is the qpa changes for emulator, but keep in mind that it works only with QEmu and not VirtualBox Mar 20 16:09:57 and I wouldn't say that I like those changes Mar 20 16:10:08 I'm not bound to VirtualBox tbh Mar 20 16:11:06 Though I never really toyed with qemu itself Mar 20 16:11:19 the problem with current QEmu support in webosose is that they never provided performance comparison with VirtualBox Mar 20 16:11:51 we have some prebuild QEmu binaries which include virglrenderer support which is supposed to be faster than old VirtualBox vboxvideo Mar 20 16:12:18 but building qemu-native with gl enabled crashes in runtime Mar 20 16:12:39 and the prebuilt QEmu isn't available e.g. for qemux86-64 Mar 20 16:12:50 and there is no integration with runqemu scripts from oe-core Mar 20 16:13:37 so in summary it's a bit more complicated to use than VirtualBox and the promissed performance isn't really verified AFAIK Mar 20 16:14:00 maybe this will get resolved before the qemux86 build is officially supported in webosose, but I doubt it Mar 20 16:26:09 From github.com:webOS-ports/meta-webos-ports Mar 20 16:26:10 e9c2ff1da..df3d2d5d3 master -> wp/master Mar 20 16:27:46 looks like Tofe's "Merge pull request #283 from Tofee/pyro-halium" went to master branch instead of pyro Mar 20 16:28:08 I'll fix that.. Mar 20 16:30:45 there is also small gcc upgrade in oe-core I'll bump that as well and retrigger the builds Mar 20 16:31:39 JaMa: https://bpaste.net/show/b3d2f453e18a Mar 20 16:32:34 https://bpaste.net/show/cb36bbe8ff5f Mar 20 16:37:32 you're right, I'm sorry Mar 20 16:38:03 there is also build time dependency on aval-impl and the only available provider for it is for qemux86 avoutput-adaptation-layer-mock Mar 20 16:38:35 and without aval-impl build time dependency it might fail Mar 20 16:38:55 So I should remove aval-impl from depends? Mar 20 16:38:57 Or. .... ? Mar 20 16:39:25 or disable that umediaserver PACKAGECONFIG for qemux86 Mar 20 16:39:50 PACKAGECONFIG_remove_pn-umediaserver_qemux86 = "avoutputd" Mar 20 16:39:54 or something like this Mar 20 16:40:56 or just PACKAGECONFIG_qemux86 = "" inside umediaserver.bb Mar 20 16:42:46 OK will try that Mar 21 01:31:26 hmm. sadly, the new build-webos in webosose doesn't run at *all* in wsl Mar 21 01:31:30 it almost instantly gives up **** ENDING LOGGING AT Wed Mar 21 03:00:02 2018