**** BEGIN LOGGING AT Wed Aug 31 02:59:58 2016 Aug 31 05:12:17 morning Aug 31 05:41:50 Tofe: Found a small bug in your refactored code while in fullscreen Aug 31 05:42:03 Half of the navigationbar is visible on top Aug 31 05:42:13 Not sure where that comes from Aug 31 05:44:34 It seems to be 1.6 Gu Aug 31 05:45:06 So I suspect https://github.com/webOS-ports/org.webosports.app.browser/blob/master/qml/NavigationBar.qml#L190 Aug 31 05:46:40 I could add a topMargin here which sets it to Units.gu(-1.6) when in fullScreen, but not sure that's the proper solution? https://github.com/webOS-ports/org.webosports.app.browser/blob/master/qml/BrowserWindow.qml#L141 Aug 31 05:49:39 That fix does work though, just not sure it's the right way to solve it Aug 31 06:57:00 Herrie: using negative length looks like a hack, there must be a better way, let me have a look Aug 31 06:59:53 Well you could do https://github.com/webOS-ports/org.webosports.app.browser/blob/master/qml/NavigationBar.qml#L190 conditionally on fullScreen Aug 31 07:13:30 yes, but I don't really understand the issue; do you happen to have a screenshot? Aug 31 07:13:59 (I can't result the browser on my job's computer, WebEngine is only for VS2013 and VS2015 and I don't have them) Aug 31 07:14:05 +test the Aug 31 07:17:03 Tofe: The bottom Units.gu(1.6) of the NavigationBar are visible in fullscreen mode for example YouTube. Aug 31 07:17:10 It'll show on top of screen Aug 31 07:20:00 ah ok, so the issue isn't that it's not fully visible, but that it's not fully hidden Aug 31 07:21:25 I think I see another way of doing it Aug 31 07:21:28 Tofe: Yeah it should be fully hidden Aug 31 07:21:48 The height is set to 0 for fullScreen for NavigationBar Aug 31 07:22:02 So it's the 1.6 that you set somehow it seems. Aug 31 07:23:04 what we can do is not change the height of the navigation bar (which results in weird height=0 widgets everywhere), but change "anchors.top: parent.top" into "y: fullScreen ? -height : 0" and have always "height: Units.gu(5.2)" Aug 31 07:23:22 (in BrowserWindow.qml line 140) Aug 31 07:24:29 That way we can also add a "Behavior on y { NumberAnimation { duration: 200 } }" to get a nice little animation Aug 31 07:24:34 I'll try that in a bit Aug 31 07:26:41 I've tried to remove most of the "height = 0" code in the refactoring, because usually you can have a nicer solution (like using Row or Column and simply set visible=false) Aug 31 07:27:26 the carddav PR is ok for merge ? Aug 31 07:27:52 Tofe: It seems to solve the issue at least, so yes. Aug 31 07:28:13 Calendar doesn't show the newly added account right away but that might have a different cause. Aug 31 07:28:24 ok Aug 31 07:28:29 It does show up after a reboot (and probably a LNC restart) Aug 31 07:28:37 Will look into the log. Aug 31 07:28:42 But the error is gone :) Aug 31 07:29:01 Tofe: That's good to know on the height = 0 :) Aug 31 07:29:04 could an activity thing or similar, maybe we haven't solve everything in that area Aug 31 07:29:10 Will keep that in mind. Aug 31 07:29:25 Tofe: Yeah just did a quick test this morning, didn't have time to check the log yet Aug 31 07:29:48 for qtwebengine, your tests were conclusive too ? Aug 31 07:30:10 Yeah nothing odd it seems. No segfaults, browser & apps seem to work fine. Aug 31 07:30:17 ok, let's go then Aug 31 07:30:31 I didn't see much in the code that should affect us anyway except for bugfixes :P Aug 31 07:30:51 no, that was quite light Aug 31 07:49:54 Tofe: Yeah I try to keep it updated at least once a month so we don't need to do big rebases :) Aug 31 08:04:05 Tofe: I made the animation 300, 200 was too short Aug 31 08:04:27 Maybe 400 is even better. I guess should have a look at some legacy bits :P Aug 31 08:04:46 PR is in :) Aug 31 08:05:44 morphis: ping Aug 31 08:06:19 Herrie|Pre3: pong Aug 31 08:06:56 morphis: Seems you have the magic keys to the LuneOS mailing list... ka6sox doesn't remember the admin password :P Aug 31 08:07:05 puuh Aug 31 08:07:08 Any chance you could add some admins to the list? Aug 31 08:07:25 I don't Aug 31 08:07:43 or let me try Aug 31 08:08:24 for http://lists.webos-ports.org/mailman/admin/luneos-dev I don't have the pw Aug 31 08:08:35 never used it and ka6sox has set it up :-) Aug 31 08:09:32 morphis: Ah Aug 31 08:09:48 You're also no moderator that can approve messages? Aug 31 08:10:03 that is what I am checking right now Aug 31 08:10:37 no, just normal subscriber Aug 31 08:11:05 Hmmz :s Aug 31 08:11:13 somebody has to reset the pw Aug 31 08:11:17 OK I guess someone will need to do some backend hacking :P Aug 31 08:12:06 Not sure if it'll be easy to reset the pw Aug 31 08:12:12 Will leave that to ka6sox then Aug 31 08:13:59 Tofe: I guess my 300 guess was spot on :P Aug 31 08:14:17 see lunaAnimations.conf in luna-sysmgr :P Aug 31 08:35:15 hehe :) Aug 31 08:36:14 300 seems most common value LOL Aug 31 09:37:58 Tofe: Ping Aug 31 09:38:20 I'm looking into 1159. Aug 31 09:38:37 Any idea how I can see which component from QML is used for this? Aug 31 09:39:47 It seems OK on desktop just on device and emulator it's broken Aug 31 09:43:14 Can I use QML Profiler for this or is there some other trick? Aug 31 09:58:11 Seems I can replicate the issue on Qemu as well. Aug 31 09:58:47 I just have a local html + css file I pushed to qemu and loaded in browser with file://media/internal/html5input.html Aug 31 09:59:03 Nothing obvious in logs Aug 31 09:59:49 It opens a new window, but it doesn't seem to populate the calendar contents and doesn't respond to clicks Aug 31 10:33:40 Herrie|Pre3: I didn't look at it yet, but does it use https://github.com/webOS-ports/qtwebengine/tree/webOS-ports/master/src/webengine/ui ? Aug 31 10:38:41 Tofe: nope Aug 31 10:38:55 It looks exactly like the calendar on Chrome Aug 31 10:39:06 So I guess Chromium supplies it and not Qt Aug 31 10:43:00 can you email me the html+css you have for tests ? Aug 31 10:44:05 Sure will do after lunch Aug 31 10:51:40 thx Aug 31 11:50:44 * DougReeder checks in Aug 31 12:04:46 Tofe: You got mail :) Aug 31 12:05:17 Added some screenshots for you :P Aug 31 12:10:08 DougReeder: Hi :) Aug 31 12:31:54 ok, yes, that'll help :) Aug 31 12:33:06 It seems that on qemu it doesn't finish the rendering Aug 31 12:33:24 And the window for me becomes unresponsive Aug 31 12:33:44 Don't have a device at hand to test the same right now Aug 31 13:37:38 Tofe: Tested on my tethered TP, seems to produce the same as qemu Aug 31 13:40:41 I think it might be related to the QPA abstraction Aug 31 13:40:53 but it's a wild guess Aug 31 13:43:45 Tofe: I was able to confirm that the input field focus issue also happens with Enyo 2. So it's not Enyo 1 only. Aug 31 13:43:55 Not sure we established that before. Aug 31 13:44:18 Anyway probably means it's something on QT/Maliit/hwcomposer/qpa side as well :P Aug 31 13:58:23 Tofe: Seems there's a few commits in upstream qt5-qpa-hwcomposer-plugin that we might want to look into. There are only 4 in total anyway it seems since we last time rebased Aug 31 13:59:31 Some might help in solving the issue on TP after screen goes off (just going based on description now) Aug 31 14:02:58 Could be, but I wouldn't bet on it Aug 31 14:04:01 Tofe: I also doubt, but worth a try at least... Aug 31 14:05:32 We might want to bump libhybris a bit too. Not much new there though anyway. Aug 31 14:07:31 GL 3 fix + off-by-one error in prop handling & android headers might come in useful while porting some new target. Aug 31 14:14:45 Herrie|Pre3, what is your email address you use for the mailing list? Aug 31 14:16:55 ka6sox: luneos (at) herrie (dot) org Aug 31 14:18:49 okay the usual suspects are now all mods Aug 31 14:38:14 ka6sox: thnx Aug 31 16:39:50 Tofe: On bumping libhybris, any specifics I should test for? Aug 31 16:40:07 If I test Mako & TP should be sufficient or should also test Maguro? Aug 31 17:11:08 mako & TP should be sufficient Aug 31 17:27:08 Anything specific I should look for? Aug 31 17:28:32 not really Aug 31 17:29:01 create some cards, close them Aug 31 17:29:26 scroll & swipe, look at a youtube video Aug 31 17:29:34 that's quite all :) Aug 31 17:29:48 most of the times it's all or nothing Aug 31 17:33:42 Tofe: Mako seems fine :) Aug 31 17:33:59 Anyway the bump is minimal and really shouldn't have any effect Aug 31 17:35:51 btw libhybris has built-in tests, you can just run test_* Aug 31 17:40:01 nizovn: Ah cool Aug 31 17:40:10 You know where these are located? Aug 31 17:41:33 Ah /usr/bin it seems Aug 31 17:41:51 yup Aug 31 17:43:43 I get some segmentation faults on test_egl, test_egl_configs, test_glesv2, test_hwcomposer Aug 31 17:43:51 I guess I need to pass in some params for those Aug 31 17:44:04 Because my LNC works fine and the GL test apps too Aug 31 17:44:57 ok Aug 31 18:12:37 nizovn: You saw my comments on the hw kbd issue? Aug 31 18:13:09 yes Aug 31 18:14:47 nizovn: Weird issue LOL Aug 31 18:15:21 Herrie: i think it's a coincidence, the real issue appears when you open new card while keyboard is shown Aug 31 18:22:40 nizovn: Ah would need to test that Aug 31 18:41:24 Herrie: interesting, if i move https://github.com/webOS-ports/org.webosports.app.browser/blob/master/qml/NavigationBar.qml#L259 to onReleased and add Qt.inputMethod.hide() in onPressed, it seems to work.(but ugly) Aug 31 18:49:54 Oh, I missed that one... Aug 31 18:50:14 onPressed and onReleased should be avoided when possible, it's really low level Aug 31 18:50:28 here, an onClicked would be enough Aug 31 18:51:15 it must have slipped through, because I even modified the verticalAlignment to depend on onPressed... Aug 31 18:53:53 https://github.com/webOS-ports/org.webosports.app.browser/pull/61 here's what I mean Aug 31 18:56:04 I didn't add Qt.inputMethod.hide() in that PR though, because I didn't test it Aug 31 19:17:03 Tofe: Seems we have a few more to look at then :P https://github.com/search?q=org%3AwebOS-ports+onreleased&type=Code Aug 31 19:17:24 And https://github.com/search?utf8=%E2%9C%93&q=org%3AwebOS-ports+onpressed&type=Code&ref=searchresults Aug 31 19:17:25 :P Aug 31 19:19:15 some of them are justified, but yes, we should have a second look Aug 31 19:19:34 Tofe: Qt.inputMethod.hide() doesn't work in onClicked(and in onPressed too), maybe because hide() is async and needs some time Aug 31 19:20:44 nizovn: what is the issue actually ? :p Aug 31 19:23:13 if keyboard is opened and press new card button, then keyboard fails to open in that new card (url bar never gets focused) Aug 31 19:23:51 oh ok Aug 31 19:25:19 if keyboard is closed, it works, so hide() could help a little, but it's not a proper solution Aug 31 19:26:20 it's possible to focus the url bar in the new card if we click outside of the bar once (like we would to unfocus that bar) Aug 31 19:26:32 not sure what it means yet Aug 31 19:28:26 yes, this is strange too Aug 31 19:29:31 the cardshell might be to blame too Aug 31 19:29:42 maybe the new card isn't properly focused Aug 31 19:30:01 would be surprising, but not impossible Aug 31 19:31:30 yes, maybe Aug 31 20:20:13 Tofe/nizovn: I wouldn't be surprised if the cause of this is also the cause of the focus issue while switching between different input fields in Enyo 1/2 apps. Because when I hide the vkb there and then focus all is OK Aug 31 20:22:22 I mean, I hide the vkb and focus on the input that J Aug 31 20:22:35 I want to use it will start to work properly. Aug 31 20:23:02 might be related Aug 31 20:23:25 Seeing browser is qml it might be mailiit somehow. Wouldn't be a bad guess to try revert to our earlier version to see if that solves stuff Aug 31 20:23:49 Or at least check the diffs between the bumps to see if something sticks out Aug 31 20:24:31 you mean the bug wasn't always there? Aug 31 20:25:52 The Enyo 1/2 wasn't Aug 31 20:26:14 For browser cards not sure Aug 31 20:26:34 might be presage or hunspell as well Aug 31 20:26:49 Just thinking out loud now Aug 31 20:27:24 mmh Aug 31 20:28:13 We bumped so much back then, qt 5.6, qpa, hybris, maliit Aug 31 20:28:46 Could be either of them... But somehow suspect maliit Aug 31 20:30:42 I guess easiest would be to remove the SRCREV that was added in from https://github.com/webOS-ports/meta-webos-ports/commit/4798600e8184825c36ba9d050d46a87c110190e1 Aug 31 20:31:05 To test so we fall back on meta-qt5 maliit version? https://github.com/meta-qt5/meta-qt5/tree/krogoth/recipes-qt/maliit Aug 31 20:32:02 Our history seems clean to test in general: https://github.com/webOS-ports/meta-webos-ports/commits/krogoth/meta-luneos/recipes-qt/maliit Aug 31 20:33:44 yes Aug 31 20:33:57 for my part I'll continue investigating this date picker Aug 31 20:34:08 Tofe: I kicked off a build Aug 31 20:34:10 looks like a window type issue Aug 31 20:34:16 So should be ready in morning I guess Aug 31 20:34:25 Tofe: This is native Chromium UI bits right? Aug 31 20:34:47 Since looks the same in QT as Chrome? Aug 31 20:40:40 Herrie: the inside of the widget is done by chromium Aug 31 20:40:52 but showing the widget is handled on qt side Aug 31 20:42:13 it's using a Toolip style currently, which we do not manage (we only manage Popup style), and i think even qtwayland doesn't handle Tooltip top-level windows Aug 31 20:42:41 I've tried using a Popup style, but failed so far Aug 31 20:43:20 So I'll quite for tonight and see how it goes tomorrow :) **** ENDING LOGGING AT Thu Sep 01 02:59:58 2016