**** BEGIN LOGGING AT Mon Mar 19 03:00:02 2018 Mar 19 05:40:32 rrch[m]: Interesting! Seems like a continuation to some extend of what was on https://github.com/openwebos/ Mar 19 05:40:52 We'd need to have a thorough look into this, but I guess some bits might be of use for sure! Thanks for linking! Mar 19 05:44:01 Morning! I'm writing it up. I will be after comments when you have looked closer. Will it at least help you with the base system, allowing development of the higher level parts of LuneOS? I see connman, wayland, QT5 etc..., but maybe that's already done for you? Mar 19 05:45:30 Seems targeted at RPi at first, with a promise to extend. The have pictures of a watch, laptop & refrigerator (I'm ignoring the rocket!) Mar 19 05:46:45 LOL Mar 19 05:47:58 Preemptive: The only "bad" thing is that they put everything up there with a single commit, so no commit history since Open webOS which makes it tedious to compare since we need to do it manually side by side and see what possible impact would be Mar 19 05:51:20 Herrie: I had a quick look at at the OpenwebOS repository - there was an update 14 days ago. https://github.com/openwebos/build-webos Mar 19 05:51:52 The website's still empty. I doubt it will come back now. Mar 19 05:54:15 Preemptive: Well consider Open webOS dead, at least that's our impression Mar 19 05:54:24 I guess this new repo is it's replacement Mar 19 05:54:35 Seems pretty much everything is there Mar 19 05:55:10 including Enact.js... Mar 19 05:55:50 Herrie: I see no Moonstone, so I assume the UI is missing, but I don;t know enough to be sure Mar 19 05:56:05 I think a big thanks should go out to EricBlade and the others who've been lobbying for this Mar 19 05:56:37 Preemptive: Well they use the same build system as we, so they could be pulling in bits from all places Mar 19 05:57:30 Herrie: I was unaware of that, but I saw EricBlade's cryptic tweet about Enyo / Enact Mar 19 06:02:58 Preemptive, EricBlade, Andolamin and others have been lobbying for open sourcing quite some time while they were @ LG SVL Mar 19 06:04:29 :-) Mar 19 06:15:54 There's for sure interesting bits there, just it will take quite some time to fully analyze Mar 19 06:16:38 But maybe some birds can help us out ;) We'll reach out to some to see if they can be of help ;) Mar 19 08:35:27 Morning! Mar 19 08:35:40 Wow, yes, quite a lot of things to process there Mar 19 08:36:23 wayland extensions, qml-webos-framework, etc... Mar 19 08:40:52 ... and no commit history...... Mar 19 08:41:39 Tofe: Yeah... And seems they use different whitespacing too.... Mar 19 08:41:50 So diffing is not that easy either Mar 19 08:41:58 Had a quick look @ luna-sysservice Mar 19 08:42:57 I'll try to see if I can reach out to some people to see if they can provide us with some more info ;) Mar 19 08:42:57 I'm looking at the wayland compositor Mar 19 08:43:14 Tofe: I guess some of the luna bits would be useful to have for sure Mar 19 08:43:21 Seems they're on QT 5.6 and Chromium 53 Mar 19 08:43:24 So that would be a downgrade Mar 19 08:43:34 Those lgsvl has Chromium 60, 63 and 64 too Mar 19 08:44:54 yes, they use older Qt... but there must be some interesting ideas. wayland extensions can be worth a look, and they also coded their own shell and surfaces Mar 19 08:45:06 (not sure why, though) Mar 19 08:46:04 Tofe: Well all the good old bits like configurator, activitymanager and luna-sysservice etc are for sure worth a look Mar 19 08:46:20 I remember morphis said that stuff needed serious update in terms of security Mar 19 08:46:32 I guess a lot of that has been covered with these bits Mar 19 08:47:16 Herrie|Laptop: Tofe: LG SVL released some more code? Mar 19 08:47:26 morphis: All it seems! Mar 19 08:47:29 or close to all! Mar 19 08:47:30 what? Mar 19 08:47:41 morphis: https://github.com/webosose and http://webosose.org/ Mar 19 08:47:43 85 repos worth! https://github.com/webosose Mar 19 08:48:01 looks like they did it again .. Mar 19 08:48:03 Out of the blue it seems.... Mar 19 08:48:24 I know EricBlade and others have been lobbying for that for quite some time while they were @ LG SVL Mar 19 08:48:41 :-) Mar 19 08:49:10 Tofe: This one might be of particular interest to you: https://github.com/webosose/wam Mar 19 08:49:19 looks fairly completish Mar 19 08:49:34 morphis: Yeah it's a lot better compared to what we had ;) Mar 19 08:49:38 I didn't build anythign yet Mar 19 08:49:41 But will try shortly Mar 19 08:50:05 Seems they have ConnMan 1.33 on board with only 2 patches, so fairly recent too Mar 19 09:01:45 QtQuickControls 1, too bad, but understandable Mar 19 10:46:22 Tofe: Didn't Quickcontrols 2 get introduced properly only in 5.7? Mar 19 10:52:49 morphis: What are your thoughts on bootd v.s. systemd for example? Mar 19 10:53:29 Herrie|Laptop: bootd is not a replacement for systemd Mar 19 10:53:48 morphis: Ah OK Mar 19 10:53:59 its used to spawn a few webOS services in the correct order Mar 19 10:54:04 Ah ok Mar 19 10:54:17 The description confused me then "Bootd is simplified upstart-like component." Mar 19 11:11:00 its more or less a workaround to support upstart events still in a systemd world Mar 19 11:12:10 morphis: Ah Mar 19 11:17:55 Well boottime isn't really an issue for LuneOS. But I can imagine that on TV it's more critical ;) Mar 19 11:43:35 Tofe: How you want to go about comparing these? Mar 19 11:44:16 I could add a new branch with the code for existing components to our GH repo and call the branch webosose or something and then we can compare them easily in place? Mar 19 11:49:11 Herrie|Laptop: I...don't know yet :) Mar 19 11:49:47 The whitespace changes don't make it easy :P Mar 19 11:49:53 Let me test 1 repo quickly Mar 19 11:52:35 https://github.com/webOS-ports/sleepd/compare/webOS-ports/master...webOS-ports:webosose?w=1 Mar 19 11:52:46 The ?w=1 ignores the whitespace differences Mar 19 11:54:33 Ok that might not have been the best exmaple :P Mar 19 11:54:37 Loads of changes there Mar 19 11:57:16 https://github.com/webOS-ports/webos-connman-adapter/compare/webOS-ports/master...webOS-ports:webosose?expand=1?w=1 Mar 19 12:00:00 Well they don't reuse Mer components like we do Mar 19 12:01:01 We don't use Mer's Connman ;) Mar 19 12:01:39 no, but we use libconnman-qt and have QML settings and LNC that use it Mar 19 12:03:23 They expose everything directly to LS2; we might want to have more something like "OS component <-> Qt/QML <-> LS2", maybe even in a generic way Mar 19 12:07:21 Tofe: Yeah the webos-connman-adapter might not have been the best example to pick anyway Mar 19 12:41:05 Tofe: This might be a better example: https://github.com/webOS-ports/filecache/compare/webOS-ports/master...webOS-ports:wosose?expand=1&w=1&B=1&Z=1&U=0 Mar 19 12:41:27 They changed braces and there doesn't seem to be an easy way to ignore those in the diff Mar 19 12:44:09 yes they reformatted all their files, unfortunately Mar 19 12:46:54 and in addition, as there is no history, there's no way to know exactly why they changed this or that Mar 19 12:47:12 It's better than nothing, but barely usable Mar 19 12:47:47 Tofe: Yeah I'll try and see if I can reach out to see if we could get this in a more usable format.... Bits in https://github.com/webosose/qml-webos-bridge are quite useful it seems :) Mar 19 12:48:03 We'll have to do a global reformatting with the same format to see if that helps Mar 19 12:48:22 We have some own implementations, but seeing these are presumably stress tested it might be worth to switch Mar 19 13:07:32 Seems I have plenty of disk space free at home Mar 19 13:07:41 I'll just try to see if their stuff builds at least Mar 19 13:13:20 I'd be curious to see how it looks like... If you get a qemu image, maybe I could download it? Mar 19 13:13:57 Tofe: Seems they only included meta-raspberry layer with RaspberryPi3 target :( Mar 19 13:14:07 Though I've seen qemu bits elsewhere Mar 19 13:14:13 Could be as easy as adding machine... Mar 19 13:14:14 Not sure Mar 19 13:14:15 yes, I've seen them too Mar 19 13:14:21 Where's JaMa when you need him :P Mar 19 13:14:53 He has no +1 that can merge meta-smartphone? Mar 19 13:15:08 Seems there were quite some updates as well that might sort our mesa issues on qemu Mar 19 13:15:08 because he seems to be more and more busy Mar 19 13:15:18 Tofe: Children tend to do that Mar 19 13:15:25 Around 1 year it's "horror" Mar 19 13:15:29 Herrie|Laptop: oh, I forgot that "detail" :) Mar 19 13:15:45 Which means, up to 6-ish months they don't move much... Mar 19 13:15:51 So you can leave them alone Mar 19 13:16:00 From then on they start to roll, turn, move Mar 19 13:16:01 and code safely Mar 19 13:16:25 hum, of course I meant to answer the previous sentence :) Mar 19 13:16:38 Then anywhere after 8-9 months they start to try to stand and crawl everywhere Mar 19 13:16:46 Which means attention is needed every second :P Mar 19 13:17:40 Or even split seconds Mar 19 13:19:35 "Don't blink." Mar 19 13:25:08 Yup ;) Mar 19 13:25:35 Tofe: Seems nyx-lib is interesting, added GPS for example :P https://github.com/webOS-ports/nyx-lib/compare/webOS-ports/master...webOS-ports:webosose?expand=1 Mar 19 13:29:02 Quite some changes but nothing too critical Mar 19 13:31:34 Quite some interesting bits added like pressure sensor, heartbeat sensors etc :P Mar 19 13:34:40 https://www.engadget.com/2018/03/19/lg-webos-open-source-edition/ Mar 19 13:34:52 well, how about that Mar 19 13:35:37 Tofe: I am curious, with LG opening WebOS, how it affects halim integration e.g? Mar 19 13:36:01 bshah: We're analyzing.... Mar 19 13:36:19 But they dumped everything for each repo in a single commit without commit history Mar 19 13:36:25 So comparing is a PITA Mar 19 13:36:29 :\ Mar 19 13:37:10 Herrie|Laptop: if there is git history somewhere, git can re-generate history to some extent Mar 19 13:38:29 there is fancy guide somewhere Mar 19 13:38:44 bshah: Well we have http://github.com/openwebos and http://github.com/webosose Mar 19 13:38:45 eric might have that. Mar 19 13:39:06 There are quite some overlapping repos, however the new one only has 1 commit Mar 19 13:39:27 In the meanwhile they changed coding style, spacing, libs etc Mar 19 13:39:38 So comparing 85 repos line for line is tedious Mar 19 13:39:45 TLDR, create empty git repo (git init), add two repos as remmote, git fetch both of them, checkout one which have history, then git merge the one of them (here I am missing one step, but it can effectively discard the hunks which are already in proper history) Mar 19 13:43:10 bshah: regarding Halium it shouldn't change much. Or maybe nothing at all. Mar 19 13:45:07 Tofe: I agree Mar 19 13:45:29 bshah: This probably means we can update some components that weren't really touched in the past and fix some bugs by updating Mar 19 13:45:37 But code changes are quite massive in some areas Mar 19 13:45:42 So we'll need time ;) Mar 19 13:45:49 okay okay Mar 19 13:46:09 Good luck guys! (well, it seems you already have good luck as you hit jackpot :p) Mar 19 13:47:04 Going through repos 1 by 1 and listing which I think might be worth and then to prioritize based on usefulness Mar 19 13:50:22 Seem to be 30-ish repos that previously existed and that we use Mar 19 13:50:31 So that'll keep us busy for a while Mar 19 13:51:59 Tofe: Local build is running Mar 19 13:52:08 Some warnings so far, but seems to progress Mar 19 14:04:57 Tofe: I guess I'll post a question on their forum ;) Mar 19 14:05:05 They seem responsive to the 2 topics so far Mar 19 14:05:11 Who knows what we'll get as answer Mar 19 14:07:59 bshah: You around? Mar 19 14:08:48 yeap Mar 19 14:12:11 Do you know what cmaction(s) is? Can be ignored similar to doze? Mar 19 14:12:37 Just tweaking setup script for Halium devices a bit so we don't need to fork all repos all the time for these changes :P Mar 19 14:22:08 bshah: ^ Mar 19 14:22:16 Tofe: Layers are a little old, but nothing too bad: https://bpaste.net/show/ea5878f86744 Mar 19 14:22:22 it is not used (cmaction) Mar 19 14:22:29 that is all I know Mar 19 14:23:30 It seems to be some gesture & pocket detection from what I could read Mar 19 14:23:35 Originating @ Motorola or so Mar 19 14:40:20 http://webosose.org/ Looks like LG is trying to embrace open source now Mar 19 14:40:30 https://github.com/webosose/ Mar 19 14:40:41 They have a guide for the rpi even Mar 19 14:42:02 lgstoian: We noticed as well Mar 19 14:42:14 Just downside is that everything got dumped without history Mar 19 14:42:33 So to compare the 30 or so components we use we need to do it line by line.... Mar 19 14:42:37 Which is a bit tedious Mar 19 14:42:49 And then of the other 55 we need to see what could be useful and what not Mar 19 14:42:59 Only rpi3 for now too, no qemux86 for example Mar 19 14:43:47 They seem to also be open to contributions. Mar 19 14:44:55 Curious to see if they plan on running this as a proper open source project where the community is fully free to bring in features and changes Mar 19 14:46:21 Curious too Mar 19 14:46:24 We welcome this Mar 19 14:46:45 We were sad to see that they stopped working on open source using http://github.com/openwebos Mar 19 14:46:55 We use about 30 components from that, updated in some way or form Mar 19 14:47:04 So now need to figure out what & how to update ;) Mar 19 14:47:20 But seeing this is tested in the field on millions of TV's that's a good thing Mar 19 14:49:58 From what I can read they are trying to expand this beyond TV's. Mar 19 14:50:10 There's talk of 'business proposals', so I'm not sure how much support there will be for projects that aren't sponsored or co-produced by LG. But perhaps I'm cynical. Mar 19 14:50:56 Also, Alan's just pipped me with an article, so now I'm grumpy! ;-) Mar 19 14:51:12 Preemptive: He tends to do that when he gets excited Mar 19 14:51:23 Probably google's fuchsia scarred them , and like samsung they want an in house OS for when Android is abandoned :) Mar 19 14:52:18 lgstoian: Yes, like Tizen Mar 19 14:52:30 lgstoian: Samsung scared them ;) Mar 19 14:52:44 LG generally follows Samsung ;) Mar 19 14:52:49 Sometimes other way around Mar 19 14:56:01 Well, I reckoned we would need a follow up article, so I'll now re-purpose mine at leisure... ;-) Mar 19 14:58:23 https://pivotce.com/2018/03/19/open-source-webos-now-brought-to-you-by-lg/ Mar 19 14:58:36 http://webosose.org/discover/getting-started/governance-model/ Mar 19 15:04:06 http://webosose.org/discover/getting-started/grant-cla/ Mar 19 15:05:27 I don't see anything about exclusivity, so maybe it's just a big corp protecting itself from 'stolen' code. Is this normal on corporate projects? Mar 19 15:12:33 This is similar to what HP had before really Mar 19 15:12:46 I think quite a lot of projects use a CLA , don't think this is unusual. I know Canonical had something like that and even KDE Mar 19 15:14:26 after some quick googling : https://en.wikipedia.org/wiki/Contributor_License_Agreement#Users Mar 19 15:31:33 I noticed there's a dev mode module... Mar 19 15:31:49 Yup Mar 19 15:34:23 Tofe: Interesting they seem to use Chromium instead of QtWebEngine wrapped Chromium Mar 19 15:36:17 https://github.com/webosose/chromium53/tree/e71abe77551664c36f50433595fbc06e895219ee/src/lg/browser Mar 19 15:38:26 This http://webosose.org/blog/updated-chromium-lagacy-wayland-support/ combined with https://github.com/lgsvl/meta-lgsvl-browser/tree/rocko/recipes-browser/chromium is also interesting Mar 19 15:45:13 JaMa: Welcome back ;) Mar 19 15:45:26 We saw some very interesting bits hitting the interwebs over the weekend ;) Mar 19 15:46:40 https://github.com/webosose/ :) Mar 19 15:47:12 Herrie|Laptop: hehe and I've seen your question in the forums :) Mar 19 15:47:32 you can build qemux86 already (and it should work in qemu), you just need to remove one component Mar 19 15:47:48 avoutput-adaptation-layer-mock Mar 19 15:47:59 JaMa: Ah OK, but we don't have machine, or just copy from LuneOS one? Mar 19 15:48:03 I saw: https://github.com/webosose/nyx-modules-qemux86 Mar 19 15:48:14 Hence I expected qemux86 to be there to some extend Mar 19 15:48:25 the machine config is from oe-core Mar 19 15:48:39 the only difference is that you will need to use QEmu instead of VirtualBox Mar 19 15:49:07 and you can ignore that error from build-webos Mar 19 15:49:20 if you set MACHINE variable to qemux86 it's OK Mar 19 15:50:20 It's not listed there, because it's not officially supported, but the recipes and components (except the avoutputd one) are still there Mar 19 15:50:38 JaMa: OK that's good to know Mar 19 15:50:52 RPi3 build is now running (useless since I don't have a RPi3 but well) Mar 19 15:50:58 Just to see if it builds Mar 19 15:56:37 :) Mar 19 16:01:28 So far it failed twice, but restarting and it continues again Mar 19 16:01:54 fetch failed or something else? Mar 19 16:02:09 there shouldn't be any missing dependencies or random issues like that Mar 19 16:06:55 JaMa: First one was boost stuff, now node-gyp Mar 19 16:07:30 https://bpaste.net/show/0317628e1b35 Mar 19 16:11:21 Herrie|Laptop: what ubuntu release you're using for this build? I guess it's not 14.04, right? Mar 19 16:13:58 16.04 Mar 19 16:14:03 Like always Mar 19 16:15:20 thanks, that's probably the reason why you're seeing this, I've already created the bug internally that it shouldn't be using nodejs from host, but it wasn't fixed before the release... Mar 19 16:15:46 Any quick fix I could try? Mar 19 16:16:26 temporary uninstalling nodejs, npm from your host is probably the fastest Mar 19 16:16:59 but you might need to cleanup ~/.node-gyp as well Mar 19 16:18:34 I don't have them installed... That's the point... Mar 19 16:18:37 https://bpaste.net/show/7e98cda34139 Mar 19 16:19:07 This nodegyp stuff is the same issue I ran into earlier I think with the nodejs webos modules when building for LuneOS Mar 19 16:21:36 strange, will need to be investigated a bit more, but still it uses ~/.node-gyp so there might be conflict in ~/.node-gyp created with host node (or in your case different version used in LuneOS builds) and node from meta-webosose Mar 19 16:22:02 it shouldn't use nodejs from hosts and shouldn't use user's ~/.node-gyp Mar 19 16:27:36 Ok so for now I'm stuck :P Mar 19 16:28:01 That node bug can go die in a fire. Mar 19 16:28:28 halfhalo: Yeah well this whole node-gyp stuff is weird :P Mar 19 16:34:44 What if you submit the LuneOS code and get LG's engineers to fix your bugs for you? ;-) Mar 19 16:35:32 LG is known for their labour saving devices! Mar 19 17:34:06 :) Mar 19 17:35:02 JaMa: when you'll have a minute, our PR for meta-smartphone is now ready for merge. I've also talked to the AsteroidOS guys about this future change, they seem ok with it Mar 19 17:35:28 JaMa: of course if you see a blocking point in what's proposed, I can work on it Mar 19 17:49:03 Herrie: https://github.com/webosose/qml-webos-bridge/blob/master/plugin/service.cpp might be useful for us, as a guide for a more complete implementation Mar 19 17:50:40 well, the whole qml-webos-bridge could be interesting to bring in Mar 19 17:51:18 but maybe some components are not really useful to LuneOS Mar 19 18:07:55 Tofe: ok will look after baby is asleep Mar 19 18:15:19 I always forget how long chromium takes to compile... I so rarely trigger yocto builds myself nowadays Mar 19 18:38:53 Tofe: The nyx changes seem fairly straight forward as in that they don't change much, just mainly add bits Mar 19 18:40:22 Same for configurator Mar 19 18:40:53 activiymanager has been split up and had hits added so wasn't easy to assess Mar 19 18:43:34 The various luna service/sysmgr bits seem tricky Mar 19 18:43:58 Would need some trial and error probably as quickest method to test Mar 19 18:48:11 Herrie|TP: yes, and we'll also need to try to summarize what has changed, on a more global level Mar 19 18:49:16 so that we can understand a bit better what we can merge, and what we don't need or can't reasonably merge Mar 19 18:50:06 the compositor work is probably one of the latter... though it can still be interesting to see what architecture choices they've made Mar 19 18:54:08 Tofe: I'd say we should aim for lower level system stuff (configurator, activitymanager, db8, filecache, ls2, luna-service etc) Mar 19 18:54:26 And then see what we could cherry pick from other bits Mar 19 18:54:46 Interesting they have own browser based on Chromium instead of qtwebengine Mar 19 18:55:09 Herrie|TP: yes, I wonder what's the reason behind this Mar 19 18:55:34 Not to depend on qt's changes in API? Mar 19 18:55:42 A layer less in between? Mar 19 18:55:48 Licensing, primarily, IIRC Mar 19 18:56:34 Licensing ? ah, the consequences of this kind of stuff never cease to surprise me :) Mar 19 18:58:24 yes, that's why there is Qt 5.6 still being used Mar 19 19:05:01 Tofe: https://github.com/shr-distribution/meta-smartphone/pull/63/commits/fe2269937952d64edd39001fe814e8ee9d57fb68 Mar 19 19:08:02 Tofe: SRC_URI is fixed in follow up commit, but please squash that part Mar 19 19:14:26 oops! let me fix that :p Mar 19 19:19:26 JaMa: done Mar 19 19:31:33 really? in pyro-halium branch? I still see /media/LuneOS in 7th and 10th commit Mar 19 19:32:23 ah it's the other recipe meta-lg/recipes-core/android-system-image/android-system-image-hammerhead.bb Mar 19 19:34:43 Tofe: ^ Mar 19 19:47:17 oops... sorry... Mar 19 19:51:24 JaMa: should be better now Mar 19 20:00:55 merged, just for rocko I had to drop those changes in removed layers (acer, asus, samsung) Mar 19 20:02:04 yes, perfect, thanks! Mar 19 20:02:26 Herrie: we're now ready to merge meta-webos-ports Mar 19 20:06:59 Tofe: DOne Mar 19 20:24:30 Herrie: then I guess it's time to start a new build Mar 19 20:28:31 Tofe: I guess so Mar 19 20:28:35 You wanna do it? Mar 19 20:28:57 ok ! Mar 19 20:30:14 done Mar 19 20:41:50 Thnx Mar 19 20:42:30 Tofe: I think we should shortly look into the users that are added via https://github.com/shr-distribution/meta-smartphone/blob/pyro/meta-android/recipes-core/android-system/android-system_1.0.bb#L35 Mar 19 20:42:37 We'd need quite some more Mar 19 20:43:02 Just I'm not really sure how the user & group adding works Mar 19 20:43:08 I guess just adding them here might work Mar 20 00:35:35 Andolamin: Is the RPi port still being updated? The links on the wiki suggest the last release was 02/17, but there are nightlies... Mar 20 02:50:15 * EricBlade busily works at porting his most recent creation to Enact **** ENDING LOGGING AT Tue Mar 20 03:00:06 2018