**** BEGIN LOGGING AT Sat Apr 27 02:59:56 2019
Apr 27 07:19:04 <Herrie|Laptop> Morning!
Apr 27 07:20:00 <Tofe> Morning!
Apr 27 07:29:27 <JaMa> morning
Apr 27 10:04:41 <Tofe> https://github.com/webOS-ports/luna-next/pull/141 tricky analysis, easy fix... usual stuff :)
Apr 27 10:05:51 <Tofe> ^ Note that I'm going back to default AutomaticFocus here; I didn't see any regression on the vkb disappearing with XdgShell. Some more testing would be needed to confirm that...
Apr 27 10:24:55 <Tofe> There was also a minor fix for JustType, I PR'ed it too
Apr 27 10:28:07 <Tofe> JaMa: https://github.com/webOS-ports/meta-webos-ports/blob/warrior/meta-luneui/recipes-qt/qt5/qtwayland_git.bbappend#L9 <-- was that imported mistakenly from mainline hammerhead branch?
Apr 27 10:55:56 <JaMa> it was in https://github.com/webOS-ports/meta-webos-ports/commit/97dcff92e7668a05bb64a8c8fba0361c97588235#diff-07d9d6744128c280693df5c830dcc607 possibly bad rebase from mainline, but it's set to empty for all AFAIK, so it shouldn't break hammerhead?
Apr 27 10:56:19 <JaMa> but I see that I never go to the "need to find out why" part
Apr 27 10:56:48 <Tofe> I just got worried because this kind of line is only done for mainline kernels, outside of hammerhead
Apr 27 10:58:05 <Tofe> for me, the "110" build of qemux86-64 works well
Apr 27 10:58:12 <Tofe> must be a VirtualBox thing then
Apr 27 10:58:41 <JaMa> what version do you have?
Apr 27 10:58:43 <Tofe> VirtualBox 6.0.6 here
Apr 27 10:59:11 <Tofe> I do have "USB tablet" input device
Apr 27 11:01:02 <Tofe> so I don't really know what's going on...
Apr 27 11:02:07 <JaMa> 110 image also clicks only in top left corner for me (like the previous build before vboxtouch SRCREV bump)
Apr 27 11:02:23 <JaMa> 6.0.6_Gentoo r129722
Apr 27 11:03:03 <JaMa> freshly imported appliance (USB tablet is there)
Apr 27 11:03:21 <Tofe> ah, I didn't go through appliance, I just attach the disk
Apr 27 11:03:34 <Tofe> I can do the appliance
Apr 27 11:03:43 <JaMa> ah, then please import appliance and we can compare the config
Apr 27 11:04:22 <JaMa> the appliance config is "legacy" for vbox-6 if you let vbox update it and refresh appliance config in meta-webos-ports there is quite a few changes
Apr 27 11:04:46 <JaMa> so your VM config can be quite different than how vbox-6 imports old appliance now
Apr 27 11:05:00 <Tofe> ok
Apr 27 11:05:41 <JaMa> as we're all now using vbox-6 it might make more sense to update the appliance config in meta-webos-ports now (I had the commit with that somewhere, but then the appliance wasn't compatible with vbox-5 so I never merged it)
Apr 27 11:05:54 <Tofe> yes indeed
Apr 27 11:06:06 <Tofe> ok, problem reproduced
Apr 27 11:06:38 <Tofe> should be easy now, let me compare
Apr 27 11:09:20 <Tofe> I have some potentially impacting settings
Apr 27 11:10:24 <Tofe> https://imgur.com/e4egFlV.png left: not working; right: working
Apr 27 11:10:57 <Tofe> so either the USB controller, or the accelerated display
Apr 27 11:13:28 <JaMa> or the chipset (not sure what is the default)
Apr 27 11:14:44 <Tofe> 3D acceleration was it
Apr 27 11:15:07 <Tofe> Damn, I really don't want to know why :D
Apr 27 11:15:39 <JaMa> hehe
Apr 27 11:15:48 <Tofe> JaMa: do you confirm mouse works better with 3D accel on ?
Apr 27 11:15:56 <JaMa> I'm looking in vbox UI to find where did I convert and export the updated config
Apr 27 11:15:59 <JaMa> sec
Apr 27 11:16:16 <Tofe> I guess it has to do with some EGL pointer extension, somehow
Apr 27 11:16:36 <JaMa> exporting the VM takes long, because .ova includes the image as well
Apr 27 11:17:04 <Tofe> I have to go to the grocery store, take your time :)
Apr 27 11:17:05 <JaMa> I think before I was exporting just .ovf somehow
Apr 27 11:18:52 <Tofe> it's already quite a good day anyway :) oh and should I rebase&merge my meta-webos-ports PR ? the luna-next* PR are already merged
Apr 27 11:19:23 <JaMa> 3D accel helps, but now the clicks are in lower right corner
Apr 27 11:19:33 <Tofe> eww
Apr 27 11:19:39 <JaMa> not directly in corner, some 100px from right and bottom
Apr 27 11:21:13 <JaMa> and the possition is slighly different on next boot
Apr 27 11:21:31 <Tofe> same here, it seems a bit random...
Apr 27 11:21:53 <Tofe> now I have it also on top-left, even with 3D accel
Apr 27 11:25:28 <Tofe> https://paste2.org/BNjvBz40 the working one on my side (every single start works well)
Apr 27 12:06:37 <JaMa> Tofe: can you try to export working one with vboxmanage export "a5a556f3-61e5-426d-b008-d9185a544d3d" -o luneos-qemux86-2.0.ovf --ovf20
Apr 27 12:06:49 <JaMa> update the machine id with the working one from vboxmanage list vms
Apr 27 12:09:08 <JaMa> I'm importing your working VM to see if it works here
Apr 27 12:10:46 <JaMa> yup your version works
Apr 27 12:13:23 <JaMa> so I can export it myself now
Apr 27 13:49:02 <Herrie|Phone> Tofe: Good stuff on the luna-next & cardshell. This fixed VKB and JustType right?
Apr 27 13:49:17 <Herrie|Phone> I noticed JustType was acting up a bit as well earlier
Apr 27 13:49:30 <Herrie|Phone> VirtualBox issue also seems close to solved?
Apr 27 13:49:46 <Tofe> yes, vkb issues should be history now.
Apr 27 13:50:18 <Tofe> For VBox just updating the appliance might do the trick, though I didn't understand why mine worked :p
Apr 27 13:51:36 <Herrie|Phone> Nice :) That solved quite some of the remaining blockers then :)
Apr 27 13:51:38 <Tofe> Herrie|Phone: though we should double-check, once we have quasi-definitive release builds, that noone can reproduce the initial vkb issue where it disappeared instantly
Apr 27 13:52:20 <Herrie|Phone> I noticed some WPA supplicant bits in the log that might be related to the WiFi status while connecting
Apr 27 13:53:00 <Herrie|Phone> Could be our wrapper needs an API update or something
Apr 27 13:53:27 <Tofe> also, something else, for the future: I think we do the card surface resize in a wrong way, leading to useless updates of the app window
Apr 27 13:54:32 <Tofe> when playing with cards, there are only few cases where we really need to resize the app itself; otherwise, just a scaling on the compositor side should be enough, no need to redraw the app
Apr 27 13:55:30 <Tofe> but that's just optimization, not really a bug
Apr 27 14:01:59 <Herrie|Phone> Could also be that our connman is not fully compatible with Mer's QtConnman bits
Apr 27 14:02:59 <Herrie|Phone> Yeah I guess that would help for performance
Apr 27 15:34:51 <JaMa> Tofe: I've changed pretty much everything in vbox UI to get as close to your working config as possible and still it doesn't work
Apr 27 15:35:09 <JaMa> last change is this version attribute which I'm not clear from where it came:
Apr 27 15:35:12 <JaMa> -    <vbox:Machine ovf:required="false" version="1.14-linux" uuid="{da17cadd-90b4-4ef2-be80-7eea1ff10a96}" name="LuneOS-qemux86-64 Clone" OSType="Linux_64" snapshotFolder="Snapshots" lastStateChange="2019-04-27T14:55:06Z">
Apr 27 15:35:16 <JaMa> +    <vbox:Machine ovf:required="false" version="1.15-linux" uuid="{5a45c65e-4d98-4cf7-ba57-0960e75eb13f}" name="LuneOS-qemux86-64-tofee" OSType="Linux_64" snapshotFolder="Snapshots" lastStateChange="2019-04-27T12:11:18Z">
Apr 27 15:37:03 <JaMa> at least when comparing exported ovf:version="1.0" the .vbox file might contain few more differences like ovf-2.0 does
Apr 27 15:41:21 <JaMa> if I create new VM from scratch I get 1.16-linux
Apr 27 15:41:32 <JaMa> lets see if it work then
Apr 27 15:42:33 <Tofe> my best bet is the way the usb controller is declared... and maybe the version influences that
Apr 27 15:47:05 <JaMa> I'm down to this diff: http://dpaste.com/0Y6N26B
Apr 27 15:47:12 <JaMa> and still yours works and my doesn't
Apr 27 15:47:38 <JaMa> yes, I've checked USB controller as well, but you cannot use EHCI/XHCI without the oracle extension
Apr 27 15:47:50 <JaMa> so in the end in both configs we were using OHCI
Apr 27 15:55:19 <JaMa> in manually created new VM it works as well, wth
Apr 27 15:56:41 <JaMa> hmm I think it's some timing issue, if you do first click after a while the VM is booted it seems to work
Apr 27 17:53:20 <Tofe> It didn't seem to be random with my config
Apr 27 17:54:20 <Tofe> but it's true I got the other config (from testing-110 appliance) to work once, when I thought it was the 3D acceleration that did the trick
**** ENDING LOGGING AT Sun Apr 28 02:59:57 2019