**** BEGIN LOGGING AT Thu Nov 26 03:00:43 2015 Nov 26 04:37:43 Andolamin, I think startDiscovery is redundant with listDevices, also. Nov 26 04:40:21 Hmm. You think discovery should start automatically when another service call happens? Yeah, that makes sense. Most of the API is modeled after ConnectSDK, so I included it because they did. Nov 26 04:41:31 I’m thinking all an app cares about is “what devices exist”. Nov 26 04:42:29 Subscriptions reduce the number of calls needed. Nov 26 04:43:17 IIRC, the Connect SDK can’t make use of subscriptions. Nov 26 04:44:18 It’s worth considering what the semantics of playing a still image in a playlist are. Nov 26 04:44:34 They do have subscriptions, actually. Nov 26 04:45:15 Specifically, should it be displayed for a fixed time, or until controls/fastForward is sent. Nov 26 04:45:30 But, if we remove startDiscovery, then we need to also require subscribe for listDevices, too, because there would be no way to return anything immediately. Nov 26 04:46:08 Right. Nov 26 04:47:39 Hmm. Probably should include a flag for that. Currently, playlist support for images on webOS isn't working properly (there's a bug in the implementation I still need to fix) and doesn't work at all on Chromecast, so that behavior isn't 100% defined. The service is going to need to abstract some of the capabilities for Chromecast, anyway. Nov 26 04:50:17 You might add a parameter to playMedia: # seconds to display still images. Nov 26 04:50:54 0, null or missing parameter would indicate indefinitly. Nov 26 04:51:21 Yeah, agreed. Nov 26 05:05:53 Seprate from the API, your service should register for the “open” Intent, with MIME Type application/x-mpegURL, if we implement Intents. Nov 26 05:07:22 It would be cool to browse to a playlist, and cast it to a TV, without the handheld needing to do the streaming. Nov 26 05:13:26 So, your API looks good. Nov 26 05:15:22 If we ever get around to adding Zeroconf, playMedia should probably also advertise the playlist URL. Nov 26 05:25:16 I gotta reboot. Back soone Nov 26 05:25:49 Both good ideas. Definitely like the mimeType handler. Nov 26 05:27:30 Sorry, I think you said something just before I rebooted? Nov 26 06:28:56 morning Nov 26 06:29:01 morphis: You around? Nov 26 06:29:11 Herrie: yeah :) Nov 26 06:29:18 sorry didn't had time to respond yesterday Nov 26 06:29:25 these days are just crazy :) Nov 26 06:30:06 morphis: Know that feeling ;) Got a little one who persists on waking up at 4:30 lately :S Nov 26 06:30:20 So I'm in bed early, not too much time to be around here a lot lately :( Nov 26 06:30:31 hehe Nov 26 06:30:43 will get funny times when I get into that situation too Nov 26 06:34:05 Hey, if you guys all do 4:00AM meetings, I won't be able to attend them :) Nov 26 06:36:18 So I'm in bed early, not too much time to be around here a lot lately :( Nov 26 06:36:20 lol Nov 26 06:37:03 Tofe: Or you join the club :P Nov 26 06:37:59 :D Nov 26 06:39:15 morphis: You remember if you found any clues with regards to the display freezing up after a while? Seems to happen on all targets I tested (Mako, Maguro, TP). Nov 26 06:39:38 Seems it's slowing down quite a bit before that, so might be a memory leak or something... Or dbus getting flooded somehow? Nov 26 06:40:12 Herrie: did you switch the android layer? Nov 26 06:40:26 morphis: No Nov 26 06:40:38 But it happens on Andolamin's Pi as well which doesn't use Android layer Nov 26 06:40:55 but you can still get in with adb? Nov 26 06:41:03 Yeah adb is fine Nov 26 06:41:12 Just display freezes either in on or off state Nov 26 06:42:34 Usually I kill luna-next when that happens Nov 26 06:43:09 Herrie: check with top if something is consuming too much memory Nov 26 06:43:35 Some time ago I concluded it was some kind of dbus mutex lock between two threads or two processes, and I thoutht Qt 5.5 (with its fixes) would solve it Nov 26 06:43:55 morphis: No top seems normal actually Nov 26 06:44:01 otherwise it's probably locking issue Nov 26 06:44:09 what did change recently? Nov 26 06:44:19 It's been around for ages Nov 26 06:44:45 I mean 1yr + for sure Nov 26 06:45:19 I often see some connman-qt5 in the backtrace Nov 26 06:45:50 but it may just be that it uses dbus more than other parts Nov 26 06:46:09 Yeah I often also see some connman activity in journalctl sometime around when my display said it freezed Nov 26 06:46:43 We can run dbus-monitor right? Nov 26 06:46:54 Any specific params or just run it? Nov 26 06:46:59 I think so Nov 26 06:47:04 gotta leave, be back in an hour or so Nov 26 06:47:07 I'll just let it run while I'm in the office and it will occur Nov 26 06:55:00 Tofe, Herrie: could be also a high dbus traffic, see https://bugs.launchpad.net/bugs/1480877 for a similar problem Nov 26 07:20:50 morphis: I'll let dbus-monitor run while in the office, quite sure it'll be stuck by the time I get back. It's usually about 1-2 hrs I think Nov 26 07:21:02 So far dbus-monitor is quiet Nov 26 07:21:17 Herrie: be sure you watch system and session bus Nov 26 07:23:24 Ah was already wondering why I don't see anything, seems it does session by default Nov 26 07:24:15 System shows a lot more ;) Nov 26 08:37:20 Herrie|Veer: 1-2 hours ? you're lucky, I get it after like 15min... Nov 26 08:38:06 But I have an outdated SIM card plugged in, so maybe connman try harder to find a 2G network Nov 26 08:45:57 Tofe: I have no sim in mine. I'm not checking actually... Nov 26 08:46:05 I just notice it happens Nov 26 19:15:50 mmh where can I find webosjs-0.1.0 source code ? Nov 26 19:45:58 Tofe: webosjs-0.1.0? What's that? Doesn't ring a bell to me Nov 26 19:47:14 Tofe: THis one? https://github.com/enyojs/enyo-webos/tree/2.5.1-release/assets/webOSjs-0.1.0 Nov 26 19:53:33 You probably need to run JSBeautifier on it though Nov 26 19:53:36 Seems minified Nov 26 19:59:38 Herrie: code rework in QtCreator does it well, too Nov 26 19:59:45 but I found what I looked for Nov 26 19:59:52 there are two mistakes in Contacts Nov 26 20:01:07 (in source/views/views.js) Nov 26 20:01:10 1. the event is onwebOSRelaunch, not onrelaunch Nov 26 20:01:35 2. it is not "e" that contains the parameters, but "e.detail" Nov 26 20:02:50 Tofe: Ah Nov 26 20:02:51 with that in mind, tadaaa Nov 26 20:02:56 So that's whwy it doesn't work? Nov 26 20:03:06 This is the new Enyo 2 contacts app right? Nov 26 20:03:11 right Nov 26 20:03:16 Ah Nov 26 20:03:25 So relaunch on Enyo 1 should work OK already? Nov 26 20:03:33 I think so Nov 26 20:03:42 Well no ;) Nov 26 20:03:47 THough I need to check Nov 26 20:04:00 If I relaunch with app menu param it only focuses for me Nov 26 20:04:06 Doesn't show app menu right away Nov 26 20:04:10 Need to check that on 3.0.x Nov 26 20:04:21 Just revived a spare TP :P Nov 26 20:04:32 Got 2 "broken" ones for EUR 50 in good state for the rest Nov 26 20:04:50 1 I was able to fix with TpDebrick and is now being flashed to webOS :P Nov 26 20:04:52 ah, but maybe my isActive fix is incomplete, it's mostly unrelated to the relaunch newContact for Contacts Nov 26 20:04:53 Had CM on it Nov 26 20:05:07 So battery ran dead Nov 26 20:05:43 Other one might be hard bricked, but can always salvage display + displaycontroller and jack port to fix up my TP4G and another TP with broken jack :P Nov 26 20:05:50 So good EUR 50 investment LOL Nov 26 20:05:56 :p yes! Nov 26 20:05:59 32GB TP + 2 TP's to be fixed with spares LOL Nov 26 20:06:43 So, should I simply PR Contacts ? Nov 26 20:08:26 I guess so Nov 26 20:08:30 DougReader wrote most of it Nov 26 20:08:55 When TP4G works properly (now touch is broken on half screen) I can debug oFono issues a bit better there as well again :) Nov 26 20:09:31 I've updated the issue comments, maybe DougReeder has an opinion about it Nov 26 20:10:29 But I'm surprised it worked before Nov 26 20:10:47 with same Enyo and webosjs versions Nov 26 20:11:14 It probably didný Nov 26 20:11:18 't Nov 26 20:12:19 Just whole relaunch was a bit broken Nov 26 20:12:26 So I guess he assumed that was the cause Nov 26 20:15:53 Tofe: On 3.0.x. when I do the luna-send -n 1 luna://com.palm.applicationManager/launch '{"id":"com.palm.app.clock", "params":{"palm-command":"open-app-menu"}}' it also just focuses app, doesn't show app menu when the app is not focused Nov 26 20:15:58 Let me check the new contact one Nov 26 20:18:34 https://github.com/webOS-ports/org.webosports.app.contacts/pull/17 <-- here's my proposal Nov 26 20:20:59 Herrie: looks like the menu is only opened once the app is focused, it doesn't chain the focusing and the opening of the menu Nov 26 20:27:18 It does chain for luna-send -n 1 palm://com.palm.applicationManager/open '{"id": "com.palm.app.contacts", "params": {"launchType": "newContact", "contact": {"nickname": "Johnny"}}}' on 3.0.5 Nov 26 20:27:33 Putting legacy contacts app on N4 now and see what it does there with latest nightly Nov 26 20:27:46 Yes, I didn't say we did it right :) Nov 26 20:28:04 Tofe: App menu behavior is like legacy Nov 26 20:28:06 So that's OK Nov 26 20:28:24 oh, we got appmenu right ? cool Nov 26 20:30:19 for contacts, it shows the new contact card, but it doesn't focus the app Nov 26 20:30:50 probably a little thing missing on our side Nov 26 20:31:38 Legacy one does for me Nov 26 20:31:48 So it goes to the app, add new contact & focuses Nov 26 20:32:05 Ok. It's not ideal, imho but it has become a more minor bug; I'd like now to work on the deadlock issue with dbus (well, probably dbus) Nov 26 20:32:05 OWO contacts is broken though for me on N4 it seems Nov 26 20:32:38 I'll debug the OWO Contacts app a bit Nov 26 20:32:46 See if WebInspector shows something etc Nov 26 20:33:05 ok Nov 26 20:33:20 btw, did you get an interesting dbus trace? Nov 26 20:34:59 Tofe: it doesn't really show much useful Nov 26 20:35:00 https://bpaste.net/show/171c82416fc2 Nov 26 20:35:29 I guess I should've output it to file instead of my adb shell in windows which runs out of lines quickly :S Nov 26 20:36:49 Hmmz this revived TP has no AUdio, I guess I need to flash the HP.tar as well with NDI, since it has some audiod-config stuff in there etc Doing that now Nov 26 20:36:53 Other than that it's all fine ;D Nov 26 20:37:13 This one is 16GB though, but well ;) Nov 26 20:38:54 :) Nov 26 20:39:39 I guess me not flashing the HP.tar caused it... NDI is easier to flash now since it can bypass all the useless crap like activation and firstuse ;) Nov 26 20:39:47 Saves hassling with devicetool.jar ;) Nov 26 20:46:02 I don't know anything about all that hacky stuff Nov 26 20:56:45 Tofe: Just basically when you need to Dr you need to do some tricks now to bypass firstuse ;) Nov 26 20:56:47 THat's all Nov 26 20:56:54 You can meta-doctor of course as well Nov 26 21:04:33 DougReeder: YOu around? Nov 26 21:14:21 Tofe: Seems we might have some problem with some of the loadable frameworks not being available at the right time Nov 26 21:15:06 This causes legacy contacts not to run and seems also the cause for the email app problem? Nov 26 21:15:48 "Uncaught TypeError: undefined is not a function" I get for COntacts when I want to exit the First Use screen by pressing "Go". this.AppPrefs = new ContactsLib.AppPrefs(function () { Nov 26 21:16:00 Seems it cannot find the ContactsLib while it does exist Nov 26 21:59:35 Tofe: Screen froze at 16:22 see log at: https://bpaste.net/show/8bd527e378c9 Nov 26 22:00:23 See line 454-460 that might be suspicious? Nov 26 22:02:50 Could be due to ALS somehow (Ambient Light Sensor)? Nov 26 22:03:09 nizovn: I know you did something on ALS, I forgot the status... You recall? Nov 26 22:05:20 I think you probably need to look around https://github.com/webOS-ports/luna-sysmgr/blob/webOS-ports/master/Src/base/DisplayManager.cpp#l2846 Nov 27 02:14:29 Tofe, Herrie, I can look at the PR later tonight. Nov 27 02:14:39 Right now I’m only on for a few minutes. **** ENDING LOGGING AT Fri Nov 27 03:00:52 2015