**** BEGIN LOGGING AT Mon Sep 14 02:59:58 2015 Sep 14 06:38:00 Morning!* Sep 14 06:49:45 Herrie: I have great hopes in Qt 5.5 for WebEngine, QDbus deadlock fixes, and maybe even other things Sep 14 07:11:44 morning Sep 14 07:15:12 Tofe: We'll see :) We have QDbus deadlocks now? Sep 14 07:15:33 Well QT keeps improving quite a lot with every release :D Sep 14 07:19:42 You've been testing 5.5 already? Sep 14 07:20:36 Tofe: yeah it gets time to try qtwebengine again once we upgrade to 5.5 Sep 14 07:21:23 Tofe: I already started a while ago to refactor luna-webappmanager to support QtWebkit and QtWebEngine at the same time: https://github.com/webOS-ports/luna-webappmanager/commits/webengine-support Sep 14 07:21:30 to have a smooth transition Sep 14 07:25:28 morphis: Since you around: Can you create a qtwebkit-plugins repo and give me and Tofe access? We're trying to add HTML5 notifications but need to add some qtwebkit plugin for that ;) Sep 14 07:25:56 webview experimentals allow you to enable notifications but offer now way to implement them currently :P Sep 14 07:25:58 qtwebkit plugin? Sep 14 07:26:19 from what I remember you have to implement that directly in qtwebkit Sep 14 07:27:27 Morphis: See http://github.com/Tofee/qtwebkit-plugins Sep 14 07:27:49 It will basically call dbus for handling the notifications Sep 14 07:28:16 that works already? Sep 14 07:28:32 I thought the platform plugin for QtWebkit is only for the WebKit1 variant Sep 14 07:28:36 but not for WebKit2 Sep 14 07:34:51 Herrie|Veer, Tofe: or do I remember wrong? Sep 14 07:35:15 morphis: It's supposed to work with WebKit 2.3 and version from Qt5 Sep 14 07:35:53 Herrie|Veer: webkit 2.3 is the webkit1 based variant Sep 14 07:35:56 not the webkit2 one we use .. Sep 14 07:36:11 Qt5 supports both Sep 14 07:36:33 Let me check what qupzilla uses Sep 14 07:36:50 That's where it comes from Sep 14 07:39:00 Herrie|Veer: they are using http://doc.qt.io/qt-5/qwebpage.html Sep 14 07:39:18 which comes from webkitwidges which still uses the webkit1 port Sep 14 07:39:59 Ah :s Sep 14 07:40:08 You might be right then and it won't work :s Sep 14 07:43:20 right Sep 14 07:43:28 you need to implement this directly into webkit Sep 14 07:44:06 ok, so let's just forget about it for now, we'll see how to support notifications with WebEngine then Sep 14 07:45:01 Sounds like a plan Sep 14 07:45:05 Tofe, Herrie|Veer: see https://trac.webkit.org/changeset/177073 Sep 14 07:45:36 that uses libnotify Sep 14 07:58:09 morphis: How do I get phone app to work on desktop? Just to put voicecall at same level as the phone app folder? Sep 14 07:58:20 Herrie|Veer: I think so Sep 14 07:58:27 I had that working in one of my branches Sep 14 07:59:00 Herrie|Veer: on https://github.com/webOS-ports/org.webosports.app.phone/commits/voicecall Sep 14 07:59:01 I get a org.nemomobile.voicecall not installed Sep 14 07:59:11 so maybe an option to merge that branch :) Sep 14 07:59:37 there are quite some good things in it Sep 14 07:59:45 like support for incoming calls through a dashboard window Sep 14 08:02:23 Also doesn't work. When I try to build the voicecall stuff in QtCreator I get a qofono-qt5 development package not found... Sep 14 08:05:04 really? Sep 14 08:05:12 there is a qtcreator config in it Sep 14 08:05:39 It's probably my cranky windows machine :P Sep 14 08:09:52 Herrie|Veer: you're using https://github.com/webOS-ports/org.webosports.app.phone/blob/voicecall/phoneapp.qmlproject? Sep 14 08:10:17 that has everything here: https://github.com/webOS-ports/org.webosports.app.phone/tree/voicecall/qml/test/imports/MeeGo/QOfono Sep 14 08:12:53 Yeah when I open it as project and run it I guess VoiceCallManager unavailable. Then qml/services/VoiceCallManager.qml:20 "org.nemomobile.voicecall" is not installed :P Sep 14 08:16:06 hm Sep 14 08:16:18 so just create that stub component :) Sep 14 08:17:47 Yeah I'll play a bit with it Sep 14 08:24:56 morphis: seems master is already up to date with Voicecall. Sep 14 08:25:04 oh really Sep 14 08:25:22 Yeah ;) Sep 14 08:25:58 Differences between master and voicecall are cosmetic changes I made for fonts etc Sep 14 08:26:19 Or maybe you have something unpushed locally still :P ? Sep 14 08:37:35 morphis: Can you check if we're running out of space? Seems the rsync is stuck on copying some IPK's Sep 14 09:41:06 JaMa: morning Sep 14 09:41:16 Got a few things for you :P Sep 14 09:42:54 1. Seems Jenkins gets stuck on rsync (build last week, today again). Maybe we're running out of space? Sep 14 09:43:38 2. For PulseAudio Sledges from SailfishOS suggested to bump a bit further, they had a few commits last week. Sep 14 09:44:56 3. Are you aware if there's a recipe in OE for HTTPS Everywhere? Seems that EFF provides 16000 or so XML files and also a script to convert them to a sqlite db. Might be useful for us to have in our browser :) Sep 14 10:09:00 3. I don't think so http://layers.openembedded.org/layerindex/branch/master/recipes/?q=everywhere Sep 14 10:09:09 1.+2. will look today Sep 14 10:13:53 OK I guess we'd have to write something then for it :P. I.e. Pull the xmls and run a pyhton script. We have some sample recipe for that? Sep 14 10:19:43 I don't know about any, but shouldn't be difficult Sep 14 10:21:45 Is that something you could pull together if I tell you what we need? Sep 14 10:30:14 not right now, but yes Sep 14 10:31:18 No rush :) Sometime in the next weeks would be fine :) Sep 14 10:42:34 Herrie|Veer: 1. yes, disk on milla is full, I'm cleaning some older builds Sep 14 10:43:03 Herrie|Veer: 2. I've change in jansa/fido to upgrade to 6.0.39, if it builds, I'll merge it and ask you for test on N4 with next testing build Sep 14 11:18:55 Does the wiki look a bit odd (like no proper style at all?) or is this just me? Sep 14 11:27:05 JaMa: Thnx Sep 14 11:50:40 Garfonso: Same here it seems Sep 14 11:51:07 Not sure where the wiki is hosted? Maybe on milla as well and the full disk was causing issues? Sep 14 11:53:25 Herrie|Veer: 2015-09-07.log:20:46 <+Herrie> I think you can remove anything pre 01-Jun-2015 Sep 14 11:53:43 did you mean also testing images for devices we no longer build? Sep 14 11:53:59 latest grouper image is from january Sep 14 11:54:26 2.1G qemuarm Sep 14 11:54:26 2.6G qemux86-64 Sep 14 11:54:26 3.7G maguro Sep 14 11:54:26 16G grouper Sep 14 11:54:26 36G qemux86 Sep 14 11:54:28 37G tenderloin Sep 14 11:54:31 44G mako Sep 14 11:59:29 No we have some recent Grouper Sep 14 11:59:40 Maguro hasn't been built for a while Sep 14 12:00:11 Ah seems you also removed the 4.4.3 image for Mako from November last year :P Sep 14 12:01:06 218G now Sep 14 12:01:15 yes, it's pre 01-Jun-2015 :/ Sep 14 12:01:36 JaMa: Well it wasn't very usable anyway Sep 14 12:01:48 And we can rebuild if needed :P Sep 14 12:02:40 ok :) Sep 14 12:02:45 new testing build running now Sep 14 12:03:07 Thnx :) Sep 14 12:03:16 That includes the new PulseAudio? Sep 14 12:03:38 You know if the webserver for webos-ports.org also runs on Mila? Sep 14 12:03:52 Seems we get 500 errors and 507 errors Sep 14 12:04:38 from list of running processes I don't think so Sep 14 12:04:44 Hmmz Sep 14 12:04:58 also the disk with jenkins home directory doesn't have anything wiki related afaik Sep 14 12:04:59 Ok will get ka6sox or scoutcamper to have a look at that then Sep 14 12:05:39 Herrie|Veer: it will probably have new PulseAudio, if my local build finishes in time Sep 14 12:05:55 but I'm not so sure about that load average: 30.34, 30.98, 35.75 Sep 14 12:08:57 Hmmz :s Sep 14 12:09:27 Maybe we'll need a reboot sometime? Sep 14 12:14:12 lets see wht ka6sox or scoutcamper|away say, restarting wiki/web should be enough Sep 14 13:35:18 ka6sox: ping Sep 14 13:35:36 scoutcamper|away: ping Sep 14 13:42:24 Herrie|Veer, semi-awake pong Sep 14 13:43:23 JaMa, what is broken? Sep 14 13:44:23 ka6sox: Seems wiki server for webos-ports is acting up. Web console in Firefox gives me 500 and 507 errors Sep 14 13:44:41 okay so wiki Sep 14 13:44:43 not feeds Sep 14 13:45:34 I think there's also an issue again with legacy Preware feed Sep 14 13:52:03 Garfonso: Any luck on the download manager? Sep 14 14:16:47 JaMa: I created a feature request for you in the LuneOS bug tracker for the HTTPS Everywhere stuff. Let me know when you need some more info (issue 981) Sep 14 14:19:33 Herrie|Veer: ok :) Sep 14 14:20:14 Should be as straight forward as running the python script but couldn't get it to work on my windows machine. Will test on *nix later Sep 14 14:20:29 On requirement is the lxml python add-on/module it seems. Sep 14 14:20:40 And python itself of course Sep 14 14:30:24 Herrie|Veer, are you seeing the webos-ports.org page now? Sep 14 14:32:08 ka6sox: I see it but it's plain 1994 text with 0 styling. CSS file IS there I checked Sep 14 14:32:41 okay...I'm asking a MW expert for some guidance as to whats happening. Sep 14 14:32:52 might be a few hours before we sort things out. Sep 14 14:33:03 Sure Sep 14 14:33:24 I get 500 and 407's still in my console in browser on load.php Sep 14 16:05:15 Tofe: Had any luck with the screen issue? Anything I can do at my end when the issue occurs? Sep 14 16:05:42 It seems to lock up fairly random. Sometimes it's within 30 minutes, sometimes a few hrs. Sep 14 16:06:06 morphis: ^ I recall you were looking into this as well. Any findings you could share? Sep 14 16:52:45 Herrie|Veer: didn't have time today, I have to debug that on-device, so at home Sep 14 16:53:36 Herrie|Veer: screen issue? Sep 14 16:57:57 Herrie|Veer, morphis: ah wait, the screen issue is something else, it's a lockup with DBus threads; some fixes are in Qt 5.5, it's worth wait for that Sep 14 16:59:11 I was thinking about the usb-in event that bypassing pincode lock Sep 14 17:00:42 the pin code thing must be something in the display management state machine Sep 14 17:00:50 or our lockscreen implementation Sep 14 17:01:00 agreed Sep 14 17:31:27 Herrie|Veer: I'm making very slow progress, because of lack of time. Currently I'm a bit occupied with my wife being pregnant. Sep 14 17:33:42 Garfonso: i thought all the work went into GETTING her pregnant, not KEEPING her pregnant? Sep 14 17:34:03 na, the most work is caring for a pregnant wife. ;) Sep 14 17:39:14 And that's several months long! Look, he doesn't even have time to work on LuneOS, poor guy :) Sep 14 17:39:21 ;) Sep 14 17:40:11 Ok, I've finally fixed back my fido build. Now let's smash this little usb-in event management. Sep 14 18:03:16 Apart from the fact that debugging on device a problem related to usb is peculiar, it seems the whole state machine has been designed so that UsbIn gets us to DisplayOn, whatever the settings are Sep 14 18:04:46 I can change that, but do you all agree? Sep 14 18:05:22 Also, I'd like to draw the state machine we have currently Sep 14 18:05:43 (a bit for fun, a bit for documentation and understanding) Sep 14 18:21:52 Ah, I just saw it's documented in the header file. Good. Sep 14 19:16:06 Tofe: I guess there's nothing wrong with USB turning on display, however it shouldn't bypass the lock Sep 14 19:16:23 It seems the code for this in LunaSysMgr was quite a bit rewritten in OWO compared to LunaCE Sep 14 19:16:57 Garfonso: Congrats ;) How long still? P.s. My wife would go to bed early, so that would give me 1,5-2 hrs each night to do stuff :P Sep 14 19:17:18 First week baby was born I wrote pretty much whole browser because we had mother in law + help around :P Sep 14 19:18:32 Herrie|Laptop: yeah it's a bit messed up currently... It's very easy to bypass the lock, with a UsbIn or Slider event Sep 14 19:19:01 Maybe have a look at LunaCE LunaSysMgr for reference? Sep 14 19:19:07 The code looked quite differnet to me Sep 14 19:19:19 Pretty much how the whole display management was done Sep 14 19:19:27 When I looked a bit at it last week Sep 14 19:19:33 I could do so yes Sep 14 19:23:04 Herrie|Laptop: no, it's still the same: https://github.com/webOS-ports/LunaCE/blob/master/Src/base/DisplayStates.cpp#L375 Sep 14 19:23:37 I can't believe it, I'm gonna try on my TP :) Sep 14 19:24:01 I think I was looking at DisplayManager.cpp last week, not 100% sure Sep 14 19:24:48 Don't remember tbh, I was looking at 2 files next to each other and they differed a lot Sep 14 19:25:44 Ok, either I'm not looking at the right code, or I misunderstood how the state machine works Sep 14 19:28:50 Let me check Sep 14 19:29:05 I think the code line I pasted was in the file that differed let me check the log Sep 14 19:31:54 I think I overlooked the updateLockState method Sep 14 19:32:50 I don't remember but I saw a significant bit of code that was different Sep 14 19:38:32 ok, well, basically they did what I wanted to do :) So it's just a little bug somewhere in that big bunch of code Sep 14 19:39:39 Tofe: https://github.com/webOS-ports/luna-sysmgr/blob/webOS-ports/master/Src/base/DisplayStates.cpp#L811 ? Sep 14 19:41:34 That basically unlocks the display when it's turned on Sep 14 19:41:40 Which I believe it shouldn't do? Sep 14 19:42:01 But then again my C++ understanding isn't that good Sep 14 19:42:56 No, that's what I thought too, but no, it's not that simple Sep 14 19:43:47 look at https://github.com/webOS-ports/luna-sysmgr/blob/webOS-ports/master/Src/base/DisplayManager.cpp#L3401, it does some checks before actually unlocking Sep 14 19:44:04 and if the checks require it, it changes the state to the Lock state Sep 14 19:44:19 it's a bit spaghetti-like, but well Sep 14 19:44:56 Yeah Sep 14 19:45:10 Well add some debugging here and there and you should get some clues? Sep 14 19:45:23 Something like that Sep 14 19:46:43 It gets quite spaghetti like :S Sep 14 19:46:53 But well there are many cases to take into account Sep 14 19:47:01 So that makes it less easy ;) Sep 14 19:47:08 :p Sep 14 19:47:20 For the display that freezes in a certain state, you expect that's DBus lock and will be solved in QT 5.5? Sep 14 19:49:51 I hope so yes Sep 14 19:50:07 It solved a great part of the issues for Ubuntu Sep 14 19:50:15 Could be it reports isOnPuck falsely as true? Sep 14 19:50:27 Hence bypasses the lock? Sep 14 19:50:43 https://github.com/webOS-ports/luna-sysmgr/blob/webOS-ports/master/Src/base/DisplayStates.cpp#L1436 Sep 14 19:50:46 Possibly Sep 14 19:52:09 isOnPuck is 'm_chargerConnected != CHARGER_NONE' Sep 14 19:52:14 Yeah Sep 14 19:52:18 That looks suspicious ;) Sep 14 19:52:24 yes... Sep 14 19:52:34 That basically would always put it on puck ;) Sep 14 19:52:40 That would also explain why it activated the DockMode after a timeout Sep 14 19:53:22 https://github.com/woce/LunaSysMgr/blob/alpha/Src/base/DisplayManager.cpp#L528 Sep 14 19:54:07 Tofe: I'm OK with the DockMode after a timeout in general Sep 14 19:54:07 I'd say https://github.com/webOS-ports/luna-sysmgr/blob/webOS-ports/master/Src/base/DisplayStates.cpp#L1048 Sep 14 19:54:41 I'm ok too, but not from Off to On bypassing lock :) Sep 14 19:54:48 Yeah ;) Sep 14 19:55:33 For the link I provided, I think the reasoning is suspicious Sep 14 19:55:59 We are in the On state, we ask for the Lock screen, and still it wonder if it should switch to it or not... Sep 14 19:57:36 Tofe: I think this is more suspicious: https://github.com/woce/LunaSysMgr/blob/alpha/Src/base/DisplayManager.cpp#L528 vs https://github.com/webOS-ports/luna-sysmgr/blob/webOS-ports/master/Src/base/DisplayManager.cpp#L508 Sep 14 19:57:48 That basically reports the onPuck incorrectly for us Sep 14 19:57:54 Legacy seems to have more proper handling Sep 14 19:58:25 Well, if you want a dockMode when usb is plugged, you have to include it as a "dock" Sep 14 19:59:08 That's probably true Sep 14 19:59:19 I didn't look too much into the rest of the code, you did ;) Sep 14 19:59:57 I think I got the issue Sep 14 20:00:26 https://github.com/webOS-ports/luna-sysmgr/blob/webOS-ports/master/Src/base/DisplayStates.cpp#L1722 Sep 14 20:00:54 same sort of reasoning: we are on Puck, we ask for Lock --> we hop to DockMode... but it's not what was requested! Sep 14 20:06:48 I understand you don't want your lockscreen to bother too much, but for me DockMode is a dimmed DockOn mode Sep 14 20:07:09 Docking a tablet to bypass any passcode is fishy Sep 14 20:07:55 Tofe: Yeah it is :P Sep 14 20:07:59 It shouldn't bypass it Sep 14 20:12:53 JaMa: Any luck with the PulseAudio? Sep 14 20:24:43 Herrie|Laptop: tempted to try an ipk ? :) Sep 14 20:25:01 Sure Sep 14 20:25:51 https://dl.dropboxusercontent.com/u/4679068/luna-next/luna-sysmgr_3.0.0-3%2Bgit1%2Be99f658af8-r0.7_armv7a-vfp-neon.ipk Sep 14 20:26:43 For me it seems to work; please do try to bypass the lock :) Sep 14 20:27:18 Ah, mmh, there is now a lock between DockMode and Puck... tsss Sep 14 20:28:03 Should it? Sep 14 20:29:57 No, there is still some bug it seems Sep 14 20:30:50 I think the problem come from the "LockDockMode" concept Sep 14 20:31:11 I'll let you play a bit more ;) Sep 14 20:31:18 I guess that problem will be mainly on TP Sep 14 20:31:26 But shouldn't be there ;) Sep 14 20:31:54 no, it'll be also on N4, as we consider that usb charging is liike a dock mode Sep 14 20:41:30 Herrie|Laptop: nothing yet, my kernel build failed for some reason Sep 14 20:41:38 Ok, I'll finish that tomorrow Sep 14 20:41:42 gn8 Sep 14 20:41:51 JaMa: Ok Sep 14 20:41:54 Tofe: Great gn8 Sep 14 21:05:59 morphis: ping **** ENDING LOGGING AT Tue Sep 15 02:59:58 2015