**** BEGIN LOGGING AT Mon Oct 28 02:59:58 2019 Oct 28 07:21:06 Morning! Oct 28 07:45:45 Morning! Oct 28 07:47:47 I'm updating my build at home, I'll soon be able to debug a bit more this justtype issue Oct 28 08:12:22 Tofe: Nice, it shouldn't be anything major really Oct 28 08:14:32 it doesn't make is easier to find :p Oct 28 08:14:36 s/is/it/ Oct 28 08:21:09 That's true... Oct 28 12:48:33 JaMa: I pushed a minor fix in meta-qt5; it's in warrior branch, maybe I should target master instead? Oct 28 12:48:44 s/pushed/PRed/ Oct 28 12:49:24 (it's unrelated with the current issue with Qt 5.12.5) Oct 28 12:54:23 Tofe: yes, it needs to be applied in all newer branches (zeus, master) as well, but I'll do it Oct 28 12:54:52 is this blocking the release (do we need new builds) or should we fix it after? Oct 28 12:55:31 I've also seen few random build failures in libqofono recently, not sure what caused them Oct 28 13:20:09 JaMa: no it's not blocking, it's a new fix that will help a bit when ofono is started too late Oct 28 18:07:30 So, for JustType, I looked at the computed DOM, and it seems we have a 'style="width:0" ' somewhere in the app Oct 28 18:07:50 removing this faulty style fixed the appearance Oct 28 18:11:40 I'd say " var size = (Math.min(window.innerWidth, window.innerHeight) * 0.95) + 'px'; " could be what causes it in app/JustType.js Oct 28 18:11:53 so maybe window.innerWidth and window.innerHeight aren't set yet?... Oct 28 18:14:28 yup, refreshing the html from inspector fixes the UI too, because window.innerWidth is set later Oct 28 18:14:36 Tofe: That should be fairly easy to check Oct 28 18:15:11 Well it must have been something in 5.12.5 webengine then Oct 28 18:15:23 Knowing the cause it should be easier to trace Oct 28 18:15:48 Weird it always worked before Oct 28 18:16:17 I would be tempted to say the webengine commit was wrong or one of our custom patches wasn't properly migrated Oct 28 18:18:47 I mean it has always worked since earlier webkit days it seems Oct 28 18:18:59 Would be weird a point release for Qt would all of the sudden break it? Oct 28 18:20:32 yes, that would be strange Oct 28 18:20:47 Ah seems that code was introduced by ShiftyAxel, but ages ago: https://github.com/webOS-ports/luna-applauncher/commit/e497bc1e349dbe73f56eb0e95c8da8c1f7be2b28 Oct 28 18:20:47 let's inspect the Qt upgrade commit :p Oct 28 18:22:10 we could also change the JS code and introduce a margin instead (which was, I guess, the real intent) Oct 28 18:23:03 https://github.com/qt/qtwebengine/commit/e44b509bad1f06d7e257bf09e80dfb5440cdbf9e Oct 28 18:23:50 ^ this commit is cryptic, but I'm not sure it's related Oct 28 18:24:23 https://github.com/qt/qtwebengine/commit/e44b509bad1f06d7e257bf09e80dfb5440cdbf9e Oct 28 18:24:25 frankly, I think ShiftyAxel's code can be improved by setting a 5% margin-left and right Oct 28 18:24:35 https://github.com/qt/qtwebengine/commit/44303861fd116b3a279d26300147e89a0bf8121c Oct 28 18:24:56 https://github.com/qt/qtwebengine/commit/56fadb571f32b721d8b99554e6e38692009ec37f Oct 28 18:25:00 https://github.com/qt/qtwebengine/commit/ddfa2997939fc4de2993b3e0de2201eca534af61 Oct 28 18:25:08 Based on commit messages those owuld be the only candidates Oct 28 18:25:17 Didn't check them in detial, since I'm having dinner next to laptop ;) Oct 28 18:26:02 https://github.com/qt/qtwebengine/commit/44303861fd116b3a279d26300147e89a0bf8121c looks like a good candidate Oct 28 18:27:03 Anywya I'm for a quick fix that doesn't require webengine rebuild ;)\ Oct 28 18:27:38 If we anyway need to fix, I suggest to merge these 2 as well and include them: https://github.com/webOS-ports/luneos-components/pulls Oct 28 18:27:57 style="margin-left: 2%;margin-right: 2%;" works pretty well it seems; and it'll allow resize Oct 28 18:28:23 I'll propose a PR for justtype Oct 28 18:29:01 I'll also merge these two yes, I agree it's a nice to have and not dangerous Oct 28 18:30:55 Yeah and we'll show the correct Qt version LOL Oct 28 18:35:34 voilà: https://github.com/webOS-ports/luna-applauncher/pull/7 Oct 28 18:36:00 Herrie: just wait a bit, I'll test it on my build Oct 28 18:37:57 I'll test it on Qemu on my end Oct 28 18:38:04 Simply modifying the file and checking what it does Oct 28 18:41:08 works here Oct 28 18:42:04 Seems OK here too Oct 28 18:42:13 Bot in portrait and landscape when I rotate screen Oct 28 18:42:24 yep, same Oct 28 18:42:27 With the F buttons Oct 28 18:42:34 So I think we're good :) Oct 28 18:42:36 ok let's merge, bump and release :) Oct 28 18:44:26 We hold off on the sensorfw one in meta-wop still right? Oct 28 18:45:03 right, that one might break some things, I didn't test it on other device than mainline hammerhead Oct 28 18:45:19 OK Oct 28 18:45:48 Where's JaMa when you need him :P Oct 28 18:46:09 what do we need him for? Oct 28 18:47:25 Well I'm not really sure how to get all this into a new staging ;) Oct 28 18:47:30 Never really had to do that Oct 28 18:47:40 I have some ideas, but not 100% sure, don't wnat to mess it up ;) Oct 28 18:47:41 ah, neither had I Oct 28 18:49:56 OK I tried to run http://jenkins.nas-admin.org/job/LuneOS/job/luneos-stable_feeds-new-staging which didn't work Oct 28 18:50:07 Seems I need to update the manifest.json first Oct 28 18:50:12 I know how to do that though ;) Oct 28 18:53:40 hm, does the input have the same width in portrait and landscape? i think that was the idea Oct 28 18:54:24 nizovn: Well it got completely broken with qt 5.12.5 upgrade as in not usable. We can always tweak it a bit still if needed going forward Oct 28 18:54:29 At least it's working again now Oct 28 18:55:47 nizovn: It seems to have 95% now in both portrait and landscape Oct 28 18:55:48 is it qt fault, or the launcher code is not good? Oct 28 18:56:02 Seems to be timing mainly Oct 28 18:56:10 value gets initiated later, so returns 0 early on Oct 28 18:56:30 i think qt fault may impact other code as well Oct 28 18:57:02 So far I didn't see anything, not sure it's Qt's fault or fauly code from luna-applauncher. Was stuff ShiftyAxel did long time ago Oct 28 18:59:26 It seemed some specific code that ShiftyAxel added, so it could just be he didn't use the best possible solution and it just worked all the time. Oct 28 18:59:40 i would prefer to solve real problem to quick hack, but well Oct 28 19:00:23 nizovn: what is the advantage of having a fixed width ? Oct 28 19:00:55 it's too large in landscape Oct 28 19:01:07 i think, didn't see Oct 28 19:07:24 with today's screens landscape is barely usable on luneos, I think Oct 28 19:07:53 on TP it's ok Oct 28 19:07:59 cardshell doesn't look good on 16:9 ratio Oct 28 19:08:07 yes, for 4:3 it's fine Oct 28 19:09:35 we can always search for the real cause of window.innerWidth not being valued yet, but that'll take quite some time probably Oct 28 19:10:12 16:9 is a challenge but seems 4:3 is making a little bit of a comeback Oct 28 19:10:49 on tablets, yes... hopefully ! 16:9 for a tablet was a terrible mode Oct 28 19:34:43 Ok new stable builds are running Oct 28 19:36:11 And failing because of too little space :S Oct 28 19:36:56 And for this I need JaMa after all or novaldex|away Oct 28 19:50:44 back to step 1 :p Oct 28 20:27:55 Tofe: YGM Oct 28 20:33:42 yup **** ENDING LOGGING AT Tue Oct 29 03:02:49 2019