**** BEGIN LOGGING AT Thu Oct 29 02:59:59 2015
Oct 29 05:53:20 <Herrie> morning
Oct 29 05:56:15 <Andolamin> Morning
Oct 29 06:31:38 <Tofe> Morning
Oct 29 06:43:10 <Herrie> Tofe: New build didn't solve youtube sound issue, so I guess we can revert to qtmultimedia
Oct 29 06:43:19 <Herrie> With email app it seems it's a focus issue.
Oct 29 06:47:48 <Tofe> ok
Oct 29 06:48:07 <Tofe> I've fixed the banner issue, but I will have to make a more generic fix for other calls
Oct 29 06:48:22 <Tofe> I used the syntax you teached me :p
Oct 29 07:02:36 <Herrie|Veer> Tofe seems adding Google CDAv in Accounts now crashes webappmgr. But after this I can open Email. Hence I'm thinking it not working normally is due to visibility/focus or broken noWindow
Oct 29 08:20:11 <Herrie|Veer> Morning from office. Reunited with my Maguro :P
Oct 29 08:30:57 <Tofe> :)
Oct 29 08:39:12 <Herrie|Veer> Tofe: Sent you some log for C+DAV. It seems we get a hard crash now, but we got further then before at least that's how I read the log.
Oct 29 09:03:56 <Tofe> ok
Oct 29 09:05:55 <Tofe> I guess InAppBrowserExtension::open(const QString&, const QString&) is the culprit
Oct 29 09:06:13 <Tofe> I didn't look at it very closely, maybe it does something not adapted to WebEngine
Oct 29 09:11:20 <Tofe> Well, it still uses WebKit, that explains things :)
Oct 29 09:19:16 <Herrie|Veer> Tofe: LOL yes :P
Oct 29 09:19:24 <Herrie|Veer> That's no longer there :P
Oct 29 09:19:28 <Herrie|Veer> So that won't work
Oct 29 09:21:11 <Herrie|Veer> Should be fairly straight forward to migrate I guess?
Oct 29 09:21:57 <Tofe> yes, I don't see anything difficult here
Oct 29 09:22:23 <Herrie|Veer> Should just update the QML to use WebEngineView instead of WebView?
Oct 29 09:22:45 <Tofe> there's also one line to remove from the cpp of the extension
Oct 29 09:23:00 <Tofe> Surprising that it even built
Oct 29 09:23:48 <Herrie|Veer> Yeah LOL
Oct 29 09:23:52 <Herrie|Veer> You'll PR it?
Oct 29 09:23:58 <Tofe> yes, tonight
Oct 29 09:24:07 <Herrie|Veer> That's at least fairly easy to fix :)
Oct 29 09:24:13 <Tofe> yes
Oct 29 09:24:32 <Herrie|Veer> Got a nightly running with your 2 PR's from this morning
Oct 29 09:24:54 <Herrie|Veer> If you tell me what line needs to be removed from CPP I guess I could do that?
Oct 29 09:25:19 <Tofe> Oh it's just that I wanted to test it before PR'ing
Oct 29 09:26:12 <Herrie|Veer> Tofe: It can't get much more broken right :P ?
Oct 29 09:35:27 <Tofe> Right, but I'd like to avoid several PRs just for doing try&see
Oct 29 09:35:43 <Tofe> It's not that urgent
Oct 29 09:35:46 <Herrie|Veer> Tofe: Yeah true :)
Oct 29 09:35:58 <Herrie|Veer> Well release time is soon :P
Oct 29 09:36:14 <Herrie|Veer> But we got some time still or can skip altogether if needed
Oct 29 09:37:05 <Herrie|Veer> Depends a bit how stable things get. Issues don't seem to big so far but didn't stress test much yet
Oct 29 09:37:17 <Tofe> We'll see how many regressions we have, yes
Oct 29 09:37:36 <Herrie|Veer> Will have to push a bunch of Enyo 1 apps onto TP and test :)
Oct 29 09:37:58 <Tofe> Yeah I only tested with what I had directly in the image
Oct 29 09:38:02 <Herrie|Veer> Last time (with webkit) it was all pretty good and pretty much every app ran
Oct 29 09:38:16 <Tofe> (and I didn't create any account, so some things I missed)
Oct 29 09:38:18 <Herrie|Veer> So hopefully it's similar.
Oct 29 09:38:36 <Herrie|Veer> Palm apps are technically the most complicated :P
Oct 29 09:38:38 <Herrie|Veer> Usually
Oct 29 09:38:50 <Herrie|Veer> 3rd party apps use less tricks usually
Oct 29 09:39:41 <Herrie|Veer> I did notice that I didn't get a list of my emails, though it said I had 8 unread emails. Will check the logs a bit, probably some deprecated Enyo 1 CSS stuff
Oct 29 10:31:51 <Herrie|Veer> lzm: FYI Tested on my Nexus 4, boot time 30s :)
Oct 29 10:34:24 <lzm> takes ~5 seconds on virtualbox
Oct 29 10:35:11 <Herrie|Veer> lzm: Depends on how much resources you assign :P
Oct 29 10:35:37 <Herrie|Veer> But 30s is pretty decent for full mobile OS as well ;) We're not bloated so boot is pretty clean :)
Oct 29 14:50:08 <lzm> i'm having a problem starting luna-next
Oct 29 14:50:10 <lzm> WARNING: 10:43:51.877: (file:///usr/palm/luna-next/shells/card/qml/main.qml:21:1: plugin cannot be loaded for module "LunaNext.Compositor": Cannot load library /
Oct 29 14:50:13 <lzm> usr/lib/qt5/qml/LunaNext/Compositor/liblunanext-compositor-qml.so: (/usr/lib/qt5/qml/LunaNext/Compositor/liblunanext-compositor-qml.so: undefined symbol: _ZNK4luna16CompositorWindow5readyEv)
Oct 29 14:50:36 <lzm> looking at the source code, i cant find where CompositorWindow::read() is defined
Oct 29 14:50:41 <lzm> ready*
Oct 29 14:53:09 <lzm> ah, found it
Oct 29 14:53:51 <lzm> hm, no, i didnt
Oct 29 14:57:39 <Tofe> lzm: https://github.com/webOS-ports/luna-next/blob/master/plugins/compositor/compositorwindow.h#L53
Oct 29 14:59:26 <lzm> that's the declaration, not definition
Oct 29 15:00:36 <Tofe> Hey, that's right, it's missing!
Oct 29 15:01:13 <lzm> how did anything work without that?
Oct 29 15:02:18 <Tofe> The symbol was probably never called
Oct 29 15:03:37 <Herrie|Veer> lzm: That's why we need your compiler :P
Oct 29 15:03:44 <Herrie|Veer> To spot these kinds of issues :D
Oct 29 15:04:53 <lzm> i am indeed analyzing luna-next with the pareon tools
Oct 29 15:05:42 <lzm> i've found that many files don't compile due to bad c++ code
Oct 29 15:06:17 <Herrie|Veer> lzm: I'm sure there are many more issues like this. It's a bit the culprit on c(++) imho ;) You can really write poor code that compiles but has issues. It's a lot less forgiving in other languages :P
Oct 29 15:06:19 <lzm> pareon is based on clang, which is much stricter than gcc, so for instance casts from const char* to char* result in a compilation error, while in gcc it results in a warning (with -fpermissive)
Oct 29 15:06:32 <Herrie|Veer> But then again I don't program much... Just some scripting
Oct 29 15:19:20 <Herrie|Veer> morphis: ping
Oct 29 15:24:27 <Herrie|Veer> Andolamin: You around already?
Oct 29 15:26:13 <Andolamin> Yeah
Oct 29 15:27:41 <Andolamin> Herrie|Veer: what's up?
Oct 29 15:27:43 <Herrie|Veer> You have any idea if there's some way for a global override for the font? Seems our Enyo 2 apps are no longer using Prelude after we went from QT 5.4.7 to 5.5
Oct 29 15:29:43 <Andolamin> Hmm, I presume it was the default font in WebKit, since I don't believe that Enyo 2 has any font specified in the base CSS, which I think Enyo 1 did.
Oct 29 15:30:04 <Herrie|Veer> Andolamin: There are some for Onyx which are not Prelude.
Oct 29 15:30:15 <Herrie|Veer> Yeah it could be that it's WebKit bug....
Oct 29 15:32:46 <Andolamin> Looking around, it looks like some of the QtWebEngine samples load at browser user style sheet to specify some styles. Maybe we need to do something similar for apps?
Oct 29 15:38:47 <Andolamin> Or QWebEngineSettings::globalSettings()->setFontFamily(QWebEngineSettings::StandardFont, ...)
Oct 29 15:42:31 <Andolamin> http://doc.qt.io/qt-5/qwebenginesettings.html
Oct 29 15:48:15 <Andolamin> If QWebEngineSettings does what we want, that's probably the best route
Oct 29 15:50:07 <Tofe> Andolamin: but we do set the fonts: https://github.com/webOS-ports/luneos-components/blob/master/modules/LuneOS/Components/LunaWebEngineView.qml#L56
Oct 29 15:51:11 <Tofe> There must be something missing, but as the bug appeared _before_ the switch to webengine, I think we should look somewhere else
Oct 29 16:01:47 <Herrie|Veer> I could for now update the CSS for all our apps as a work-around but that won't solve it for 3rd party apps using Enyo 2 (though that number is fairly small :P)
Oct 29 16:32:56 <morphis> Herrie|Veer: pong
Oct 29 16:33:55 <Herrie|Veer> morphis: disabling qtmultimedia and enabling gstreamer again didn't sort the no sound issue on YouTube. You have any suggestions on how to debug the culprit? System sounds do work, so it's not a generic audio problem.
Oct 29 16:34:23 <morphis> Herrie|Veer: right, there is a way how chromium sends out audio
Oct 29 16:34:32 <morphis> all audio goes through pulseaudio
Oct 29 16:34:39 <morphis> in our case and must go this way
Oct 29 16:34:48 <morphis> so possible that this route isn't setup properly
Oct 29 16:35:04 <morphis> didn't looked that much into chromium yet that I can tell a easy way how to figure that out
Oct 29 16:35:30 <Herrie|Veer> Ah ok I can search a bit based on above
Oct 29 16:40:34 <Herrie|Veer> Seems it occurs elsewhere on distro's like arch too, seems I might need to play a bit with alsa, pulseaudio etc. Will have some tests tonight :)
Oct 29 16:55:39 <lzm> i'd like to run luneos on a nexus5, would it be too much work to add support for it?
Oct 29 16:58:48 <Herrie|Veer> lzm: It's fairly doable. We're reusing lots of work from CyanogenMod as a basis. Nexus 5 is fairly well supported by libhybris as well (Jolla/Sailfish OS guys use it too). So there shouldn't be too many issues getting it ported. It's all down to time and
Oct 29 16:58:48 <Herrie|Veer> resources.
Oct 29 16:59:22 <Herrie|Veer> As you can imagine we're a bit tight on that, so we try to focus on getting the system stable & bugs fixed before porting to other targets.
Oct 29 16:59:44 <Herrie|Veer> But we do have (an incomplete) porting guide to get devs started
Oct 29 16:59:47 <lzm> i would be the one doing that, in this case
Oct 29 16:59:53 <lzm> i may try that tomorrow
Oct 29 17:00:06 <lzm> do you have a link for the guide?
Oct 29 17:00:11 <Herrie|Veer> lzm: It can be found on the wiki
Oct 29 17:00:45 <Herrie|Veer> http://www.webos-ports.org/wiki/Porting_Guide
Oct 29 17:00:56 <Herrie|Veer> Any improvements are welcome.
Oct 29 17:01:28 <Herrie|Veer> I can set up your proper access in wiki once you register an account
Oct 29 17:03:49 <Herrie|Veer> Nexus 5 has good support in libhybris :) All supposedly works except for RTC alarms and NFC is unknown ;)
Oct 29 17:04:15 <Herrie|Veer> Doesn't mean we have all working in LuneOS yet though :P
Oct 29 17:05:30 <Herrie|Veer> morphis: ^ CM11 still the best source to start with?
Oct 29 17:06:27 <morphis> Herrie|Veer: yes
Oct 29 18:38:59 <Tofe> lzm: do you have an output of the compilation errors for luna-next ? I've fixed the one about ready(), but if there are other, it'd be good to fix them too
Oct 29 18:39:49 <lzm> the other compilation errors came from -fpermissive in different recipes
Oct 29 18:39:53 <lzm> like luna-appmanager
Oct 29 18:41:41 <Tofe> Herrie: ping
Oct 29 18:41:42 <Tofe> lzm: ok
Oct 29 18:47:12 <Herrie> Tofe: Banners work now
Oct 29 18:47:31 <Herrie> Seems we have issues with Units.gu for dash height, system popup
Oct 29 18:47:41 <Herrie> That probaby also explains why power menu is gone
Oct 29 18:47:57 <Tofe> Herrie: https://github.com/webOS-ports/luna-webappmanager/pull/37 will fix accounts :)
Oct 29 18:49:56 <Herrie|Laptop> Switched to laptop ;)
Oct 29 18:51:07 <Herrie|Laptop> Tofe: Nice work on the luna-webappmanager
Oct 29 18:51:22 <Herrie|Laptop> Not sure why we would use experimental for transparent bg and then set it to false ?
Oct 29 18:51:38 <Tofe> Not sure either. It worked very well without it.
Oct 29 18:52:18 <Tofe> I could maybe have increased a bit the devicePixelRatio, but that's all
Oct 29 18:52:34 <Herrie|Laptop> Tofe: To me that would mean it would set no transparent bg, so why would you bother to put it there in the first place :P
Oct 29 18:52:56 <Tofe> Also now we don't store mFrameName, which makes it more robust if you want to open multiple in-app browsers like this
Oct 29 18:53:33 <Herrie|Laptop> :)
Oct 29 18:53:44 <Tofe> mmmh power menu seems all broken, now
Oct 29 18:54:21 <Tofe> "height:  0" ah, yes, of course...
Oct 29 18:54:32 <Herrie|Laptop> Tofe: It's the GridUnits not working properly
Oct 29 18:54:39 <Herrie|Laptop> See also the ones with GridUnit in Testr
Oct 29 18:54:43 <Herrie|Laptop> They are also wrong
Oct 29 18:54:47 <Herrie|Laptop> System popup and dashboard
Oct 29 18:55:06 <Herrie|Laptop> Quite sure it's the same issue ;)
Oct 29 18:55:27 <Tofe> Do you have the piece of code opening the powermenu, by any chance ?
Oct 29 18:55:43 <Tofe> The enyo popupalert open call, I mean
Oct 29 18:56:05 <Herrie|Laptop> https://github.com/webOS-ports/luna-systemui/blob/webOS-ports/master/data/PowerdService.js#L212
Oct 29 18:56:13 <Herrie|Laptop> That's the one ;)
Oct 29 18:56:25 <Herrie|Laptop> Was just looking for it already, because I guessed you'd ask for it :P
Oct 29 18:56:35 <Tofe> 25 would be the wanted height ?
Oct 29 18:56:45 <Herrie|Laptop> 25 gridunits yes
Oct 29 18:56:49 <Tofe> ok
Oct 29 18:57:09 <Tofe> now let's look into Enyo to see how this is translated into a window.open call
Oct 29 19:00:52 <Tofe> ok, should be translated into a classic height=25 attribute
Oct 29 19:02:13 <Tofe> let's go back to debugging WebEngine
Oct 29 19:02:43 <Herrie|Laptop> It was working OK before when you fixed it in WebKit :P
Oct 29 19:02:49 <Herrie|Laptop> So it's probably something small
Oct 29 19:12:45 <Herrie|Laptop> lzm: What will you use as a build machine? Something with lots of CPU & memory will speed up things quite a bit :)
Oct 29 19:19:58 <lzm> figuring out what to do would take 10x more time than building
Oct 29 19:22:34 <Herrie|Laptop> lzm: Yeah true, but if you need to build & rebuild a beefy machine helps :)
Oct 29 19:22:49 <lzm> im not sure i'll be able to spend time adding support for the n5, i have to talk to my managers. it might be better for me to keep using the emulator
Oct 29 19:22:52 <Herrie|Laptop> I.e. 8+ GB of memory, 16 or 32 always better ;) And 4+ cores helps ;)
Oct 29 19:23:12 <lzm> i have access to a 40-core machine with 128gb of ram at work
Oct 29 19:23:17 <Herrie|Laptop> lzm: Yeah that's always the problem ;)
Oct 29 19:23:19 <lzm> building is not a problem
Oct 29 19:23:21 <Herrie|Laptop> That's nice :D
Oct 29 19:23:43 <Herrie|Laptop> Our builders are also beefy and use SSTATE CACHE across devices with same architecture ;)
Oct 29 19:23:56 <Herrie|Laptop> So usually most builds finish within 20-30 mins :P
Oct 29 19:27:34 <lzm> where does the funds for the build machines come from? who sponsors the project?
Oct 29 19:28:57 <Herrie|Laptop> lzm: ka6sox should be able to answer that. But I think we still use some of the hardware that was donated by HP some years ago when they open sourced parts of webOS ;)
Oct 29 19:29:47 <Herrie|Laptop> http://www.webosnation.com/hpgram-donates-five-super-hefty-servers-webos-ports
Oct 29 19:31:05 <Herrie|Laptop> If not all of it ;)
Oct 29 19:31:41 <lzm> nice amount of ram
Oct 29 19:32:02 <Herrie|Laptop> Other costs (hosting etc) are probably covered by donations, not 100% sure.
Oct 29 20:38:02 <Herrie|Laptop> Tofe: Any ideas so far?
Oct 29 20:56:36 <Herrie|Laptop> Tofe: I suspect the culprit is somewhere in the 20 lines following https://github.com/webOS-ports/luna-webappmanager/blob/91b0761b25fc15eeb8f802cba09aa84e371292e7/src/webapplication.cpp#L217
Oct 29 21:15:41 <Tofe> Herrie|Laptop: unfortunately it's a bit more complex than that
Oct 29 21:16:56 <Tofe> The culprit is here: https://github.com/Tofee/qtwebengine-chromium/blob/8c9615c9be72bbf82e725fde8245c53dad8f4f37/chromium/content/renderer/render_widget.cc#L1393
Oct 29 21:18:17 <Herrie|Laptop> Tofe: Ah so you use widgets for this? And it automatically assumes the same place every time?
Oct 29 21:19:35 <Tofe> well, chromium uses whatever it must, I don't have any input on this
Oct 29 21:19:58 <Tofe> but the initial position, which I thought was reliable, isn't.
Oct 29 21:22:08 <Tofe> I would have a quick&dirty way of fixing it, but I'd prefer finding the right way
Oct 29 21:23:14 <Herrie|Laptop> Tofe: Seems they renamed it a while back? https://1001-9732e57c756f-tainted-dot-chromiumcodereview.appspot.com/889403004/
Oct 29 21:24:30 <Tofe> well, a new shiny name doesn't make it more reliable :)
Oct 29 21:25:16 <Herrie|Laptop> Tofe: No but it explains why it might cause some confusion about what it does
Oct 29 22:41:45 <Herrie> Tofe: Email is still broken with latest fixes, also the account name that's coming back from Google I remember to be differently. I'll check with an older image.
Oct 29 22:41:52 <Herrie> Also app menu doesn't work in Preware
Oct 29 22:42:11 <Herrie> I didn't expect email to be fixed, it's just observation
**** ENDING LOGGING AT Fri Oct 30 02:59:59 2015