**** BEGIN LOGGING AT Tue Apr 26 02:59:56 2022 Apr 26 07:55:01 Morning! Apr 26 08:03:46 Morning! Apr 26 08:04:30 Tofe: Is the hwcomposer route the correct one to follow for the Halium devices? Apr 26 08:04:36 Or we should be using eglfs there too? Apr 26 08:04:46 I noticed that rotation works on mido when I use hwcomposer Apr 26 08:05:25 it's hwcomposer iirc yes Apr 26 08:05:42 can be confirmed with our luna-next.conf for the halium devices Apr 26 08:06:14 Yes Apr 26 08:06:27 Just I don't have touch, but that could be another variable that needs adding? Apr 26 08:06:35 I also noticed that there is an eglfs_webos ;) Apr 26 08:07:20 Which they seem to use for the auto variant of the surfacemanager: https://github.com/webosose/auto-luna-surface-manager/blob/master/startup/product.env.in#L40-L41 Apr 26 08:08:01 not sure what's the difference Apr 26 08:10:25 but it seems to require kms, i.e. won't work on halium Apr 26 08:11:28 Ah OK Apr 26 08:12:49 Just wonder what would be the best approach for getting the hwcomposer value in, change the .env somehow to make it conditional based upon existence of /sys/class/drm or to use a new .env altogether for Halium Apr 26 08:13:24 https://github.com/webosose/qtbase-plugins-webos I wonder why they bothered implementing their own stuff Apr 26 08:15:13 ah, they have their own mouse/keyboard handlers, maybe that helped them on qemu Apr 26 17:08:05 Tofe: Touch I can get working similar like with luna-next too. Adding -plugin evdevtouch:/dev/input/event1 to surface-manager start after the platform does the trick Apr 26 17:09:18 For mido that is Apr 26 17:09:26 I didn't test Tenderloin yet, but that should be the same Apr 26 17:09:49 I guess we should put it in a json config array similar to the compositorGeometry and use that Apr 26 17:12:31 Seems for VKB we might need to add the evdevkeyboard as well? Apr 26 17:12:41 There's a guide for Jetson Nano board with some pointers: https://github.com/webosose/website/blob/0444dc39da3d3ac7bb17dc9ac190ee99bbebef6a/content/en/docs/guides/setup/bring-up-guides/bringing-up-jetson-nano.md Apr 26 17:23:02 For Maliit we might need to set the QT_IM_MODULE variable? Apr 26 17:25:01 Seems that there's some global conf file which we might not have (yet) and can explain our issues with XDG_RUNTIME_DIR etc? https://github.com/webosose/webos-initscripts/blob/d117b436d73418341f228b881890ddb237510674/rootfs/etc/systemd/system.conf.d/webos-global.conf.in Apr 26 18:24:53 For Geometry seems you were right, came across this: https://github.com/webosose/luna-surfacemanager/blob/168865d95dd6d3a30af2dac73e0d7236d6ac5030/modules/weboscompositor/weboscompositorwindow.cpp#L287-L289 Apr 26 18:43:32 Tofe: Ah seems we already pass bits for qemux86(-64) but not for the other targets. I guess this would be the right place to add those: https://github.com/webOS-ports/meta-webos-ports/blob/703bdbbb0a42951082f82f6ba7fbccf1c6e38ba9/meta-luneos/recipes-webos/luna-surfacemanager/luna-surfacemanager-conf/qemux86-64/surface-manager.env Apr 26 20:40:26 yes, QT_IM_MODULE is needed, but it should be there Apr 26 22:47:52 Tofe: Ok some PR's for tonight: https://github.com/webOS-ports/luna-surfacemanager/pull/1 Apr 26 22:50:35 https://github.com/webOS-ports/meta-webos-ports/pull/517 Apr 26 22:50:50 That should get the config working on our other targets as well Apr 26 22:51:09 I'm sure there are other ways, but this seems fairly minimal in terms of patching the product.env in lsm Apr 26 22:51:31 I've tested it on mido & qemux86-64 and confirm it works as expected there Apr 26 22:51:51 Without the check for /sys/class/drm the product.env does havoc on the Halium devices **** ENDING LOGGING AT Wed Apr 27 02:59:56 2022