**** BEGIN LOGGING AT Wed Oct 19 02:59:58 2016
Oct 19 04:38:43 <Herrie> morning
Oct 19 06:54:36 <Tofe> Morning!
Oct 19 07:58:45 <Herrie|Pre3> morning
Oct 19 08:42:12 <Herrie|Pre3> Tofe: Been testing qtwayland changes from upstream, I didn't see anything odd but it might be good to thoroughly check. Some commits might be fixes for some fixes you put in our fork and might therefore be duplicate, though there were no merge conflicts.
Oct 19 08:45:01 <Herrie|Pre3> I also checked upstream qt5-qpa-hwcomposer which seems to work (they merged the 5.6 changes). Just need to confirm if we need your fix for QScreen::orientation still. If not we could follow upstream directly and drop our fork.
Oct 19 08:46:50 <Tofe> Herrie|Pre3: for the orientation, you mean in the qpa ?
Oct 19 08:47:53 <Herrie|Pre3> Tofe: Yeah
Oct 19 08:47:59 <Herrie|Pre3> Seems like we might still need that
Oct 19 08:48:14 <Herrie|Pre3> But we could do that as .patch instead of a whole fork
Oct 19 08:52:30 <Tofe> What's to be merged from upstream?
Oct 19 08:53:29 <Herrie|Pre3> Tofe: upstream qpa added support for 5.6 so we can drop the majority of our custom bits.
Oct 19 08:53:41 <Herrie|Pre3> Just the orientation we still need it seems.
Oct 19 08:54:22 <Herrie|Pre3> Not sure about the QPA Interface IID where we seem to distinguish between different QT versions and upstream doesn't care at all :P
Oct 19 09:26:35 <Tofe> For the IID, upstream's version is more generic
Oct 19 09:27:34 <Herrie|Pre3> Seems so.
Oct 19 09:28:02 <Herrie|Pre3> So I guess we'd just need the sensor/rotation stuff. Is there any way to check if stuff creates havoc?
Oct 19 09:30:08 <Tofe> For QPA, most of the time is all or nothing. Create some windows, delete them, rotate device, try the keyboard.
Oct 19 09:32:00 <Herrie|Pre3> Tofe: OK. I'll test a bit more. I hoped some of these changes would fix our input focus problem, but no luck :'(
Oct 19 09:40:49 <Herrie|Pre3> Tofe: I noticed with the new card swipe implementation we have a "loading" artifact after closing cards. It's not very noticeable on the N4 but it's quite visible on slower targets like the Maguro and probably also TP.
Oct 19 10:41:21 <Tofe> you mean we see the "Loading" luneos logo of the card for a split second?
Oct 19 10:43:22 <Herrie|Pre3> Tofe: Yeah
Oct 19 10:43:31 <Herrie|Pre3> A bit longer on slower targets
Oct 19 10:57:10 <Tofe> mmh let me see how this could happen
Oct 19 10:59:22 <Tofe> ok, it's because of this: https://github.com/webOS-ports/luna-next-cardshell/blob/7e7b2375e189b94eb73533ce9903c432e85a7732/qml/CardView/CardWindowWrapper.qml#L63
Oct 19 10:59:59 <Tofe> we unmap the window before its destruction, and the latter is sometimes not immediate
Oct 19 11:01:08 <Tofe> Herrie|Pre3: where does the loading icon appear ? in the center, overlapping another card ?
Oct 19 11:43:57 <Herrie|Pre3> Tofe: I think it's in the center need to double check.
Oct 19 11:52:39 <Herrie|Pre3> Tofe: Let me check in a bit. I guess I'll put cpu speed down in VirtualBox that should make it easier to spot :)
Oct 19 12:42:45 <Herrie|Pre3> Tofe: Seems I cannot replicate it on qemu
Oct 19 12:43:04 <Herrie|Pre3> I guess it might be something in a layer we only have on an actual device?
Oct 19 13:41:31 <Tofe> mmh it could depend on wayland's backend
Oct 19 14:13:29 <JaMa> Herrie|Pre3: I've separated qt 5.7 changes from morty branch, so now morty/unstable should be enough for testing morty
Oct 19 14:45:09 <Herrie|Pre3> JaMa: Thnx will give that a go, hopefully tonight.
Oct 19 17:03:02 <Herrie> JaMa: We need to drop https://github.com/webOS-ports/meta-webos-ports/blob/morty/meta-luneos/recipes-core/packagegroups/packagegroup-luneos-development.bb#L41 for morty
Oct 19 17:06:41 <nizovn> Herrie: what color do you prefer for links? blue on blue background really hard to distinguish
Oct 19 17:08:44 <Herrie> nizovn: Will ask our graphics guy if he has some ideas ;)
Oct 19 18:38:30 <Tofe> Herrie:  would you have, by any chance, the journal log when you have the splash icon bug for a destroyed card?
Oct 19 18:41:16 <Tofe> I'm pretty sure I can have an easy fix, but I'll need to know in what order things happen for the bug to appear
Oct 19 18:58:56 <Herrie> Tofe: Hmmz I need to have a look
Oct 19 18:59:07 <Herrie> Let me flash a fresh image without my qtwayland + qt5 qpa stuff
Oct 19 19:03:57 <Tofe> :) no rush
Oct 19 19:52:50 <Herrie> Tofe: YGM
Oct 19 19:53:01 <Herrie> I wasn't really able to visually spot it while trying this time
Oct 19 19:53:15 <Herrie> Though 1 card would act a bit weird while it would be destroyed at times
Oct 19 19:53:23 <Herrie> But I didn't really see the loading bit this time
Oct 19 19:53:34 <Herrie> Maybe you see something in the log that gives a clue (or not...)
Oct 19 20:06:49 <Tofe> oh thanks
Oct 19 20:07:17 <Tofe> I'm mainly interested in the order of the log messages
Oct 19 20:08:28 <Herrie> Tofe: I opened some apps, created some stacks and swiped them away a number of times
Oct 19 20:08:35 <Herrie> I didn't really see the visual artifact this time
Oct 19 20:11:09 <Tofe> Herrie:  not related, but I see a lot of "QFont::setPointSizeF: Point size <= 0 (0.000000), must be greater than 0" in the logs... there seems to be a coeficient of zero somewhere there
Oct 19 20:19:43 <Tofe> Herrie: https://github.com/webOS-ports/luna-next-cardshell/commit/eb10e472519d8a71b8ec2b3a2b4fd3a5a00a32e0 should to it (from tofe/fixes branch)
Oct 19 20:21:14 <Herrie> Tofe: Easy fix :)
**** ENDING LOGGING AT Thu Oct 20 02:59:59 2016