**** BEGIN LOGGING AT Tue May 22 03:00:05 2018 May 22 05:54:06 Morning! May 22 05:54:44 elvispre: Like Tofe said they're not directly related, however LG would like to have Anbox on OSE it seems and we also had it as a goal for quite some time to get it running May 22 07:13:26 Morning! May 22 10:50:17 Herrie: Ah! Good info. Thanks. May 22 11:14:26 elvispre: Just seems optimal kernel version is 4.x, reports from people getting it working on 3.10 and 3.18 versions May 22 11:18:14 Seeing we only had 3.4 kernels before, it didn't make much sense May 22 18:55:51 Herrie: I'm looking at webOS OSE compositor and qtwayland integration, and they really did it the other way around... May 22 18:57:19 They extended the wayland protocol, the qtwayland protocol, to fit in the concepts (groups of cards, carded/maximize/fullscreen, etc) May 22 18:57:33 whereas we implemented all that on the QML side of the cardshell May 22 19:14:10 Tofe: Hmmz OK... May 22 19:14:19 Any ideas why they might have done this? May 22 19:14:35 I guess you could argue one way or another might perform better? May 22 19:16:50 I guess our way is easier/more flexible and easier to extend? May 22 19:18:41 Our ways is much less work :) May 22 19:19:01 We rely on Qt and QtWayland, much more than what LG did May 22 19:19:45 It looks good, but frankly I think they overdid it a bit May 22 19:20:20 The whole goal of QtWayland is to be able to have a simple compositor with just a bunch of lines of code and/or QML May 22 19:21:20 I think (but they'll have time to explain) that a good part of their code could be dropped with 5.9. Even with 5.6. May 22 19:22:41 Hehe May 22 19:22:55 On the other hand, I also think that extending the wayland protocol to communicate with the webOS windows (are you ready, what is your appId, etc) can be a good alternative to what we do (i.e. setting properties on windows, like "_LuneOS_AppId=toto") May 22 19:24:25 I'll do a drawing to explain our own compositor more in detail May 22 19:28:02 OK :) May 22 19:28:19 Yeah could be a "future" compositor could pick best from both worlds May 22 19:28:58 I do hope so May 22 19:29:31 On our side, maybe we do too much in one big QML cardshell, and we could split things in a more elegant way May 22 19:30:49 They seem to have a more elegant MVC implementation for instance May 22 19:31:09 but, they reimplemented many things that are (now) native within qtwayland May 22 19:32:07 window state, window aggregation, parenting, this is more conveniently managed at the QML level than at the wayland level May 22 19:33:55 but we'll see. We'll tell what we have, they'll do the same, and we'll merge for the best. I hope. :) May 22 20:27:17 Yup indeed :) **** ENDING LOGGING AT Wed May 23 03:00:01 2018