**** BEGIN LOGGING AT Tue Sep 20 02:59:58 2016 Sep 20 05:39:45 moorning Sep 20 05:40:16 Herrie|2, moorning Sep 20 07:23:37 morning Sep 20 08:44:17 Morning! Sep 20 08:48:28 Tofe: Morning! Back in one piece? Sep 20 08:48:33 Been somewhere nice? Sep 20 08:56:56 Herrie|Pre3: yes, we've visited a lot of castles along the Loire Sep 20 08:57:02 was really nice Sep 20 08:57:13 Yeah can imagine :) Sep 20 08:58:37 Oh, the project page has been updated: https://www.openhub.net/p/webos-ports/ Sep 20 09:01:25 Tofe: Nice :) Sep 20 09:04:02 4 years ago, we were at 200 contributors per month ? Sep 20 09:04:19 Tofe: I doubt that really Sep 20 09:04:26 Those stats must be wrong Sep 20 09:04:38 It also takes committers from other projects. Sep 20 09:04:46 I.e. Stuff we imported from OWO Sep 20 09:04:52 yes, or from Qt Sep 20 09:05:00 Yeah :P Sep 20 09:05:08 I saw one of the QT guys as committer Sep 20 09:05:28 So I guess it took all HP/LG employees there :P Sep 20 09:05:57 ok, yes, no wonder then :p Sep 20 09:08:09 While trying to get the smsmessage db kind working on qemu (we don't have webos-telephonyd there), I found that we have quite some db kinds, permissions and activities with incorrect JSON syntax. Checked ALL of them now and updated accordingly :) Sep 20 09:08:12 Time to update my qemu build, if I want to analyze the html5 file api pb Sep 20 09:08:20 Should solve some issues here and then :P Sep 20 09:08:29 s/then/there Sep 20 09:08:59 Herrie|Pre3: that will certainly improve things yes, that's a good catch Sep 20 09:09:23 We still need to look into 839 at some point too. Seems that db kinds are created too late. Sep 20 09:09:40 So I guess systemd startup order would need some tweaking here and there. Sep 20 09:11:07 Will see if I can figure out what's wrong there. Sep 20 09:14:12 Herrie|Pre3: keep in mind that the configurator has a "first launch" mode too Sep 20 09:15:01 So until after the first reboot, we are quite sure everything isn't in place Sep 20 09:15:13 That could certainly be fixed with some systemd magic Sep 20 09:24:16 There are some things that should run even BEFORE our FirstUse app I guess. And then some that could run afterwards. Will check to see if I can map how things are currently done and how they would make sense. Sep 20 09:51:05 Tofe: Dropped you some email about the download manager integration for browser. Sep 20 09:51:42 Doing the actual download seems straight forward enough, just not sure how we handle the mime type stuff. Sep 20 09:52:20 thanks, will have a look Sep 20 10:36:04 OK found a few tweaks to be done to systemd will test tonight on a new image to see if they give desired results. Sep 20 10:37:19 mediaindexer is being started way too early should be at least after configurator so we have the db kinds. Same applies for lunasysmgr if I base on their original upstart file. For mediaindexer I don't expect issues, but lunasysmgr might be a bit trickier Sep 20 10:37:19 . Sep 20 13:02:50 Tofe: On my qemux86 it seems to behave Sep 20 13:02:58 good :) Sep 20 13:03:03 Would need to test it on an actual device as well I guess Sep 20 13:03:23 A lot less errors in the log on first boot :) Sep 20 13:03:53 great! Sep 20 13:04:01 Will update app-services recipe a bit to add some dirs as well that configurator expects but we don't have. Just makes log messy currently :P Sep 20 13:04:34 I could either remove the code, but prefer to just add the dirs for any apps relying on these deprecated paths etc Sep 20 13:06:11 Tofe: You know whats up with our cryptofs? Is it proper cryptofs or just mounted as being cryptofs but unencrypted? Sep 20 13:06:37 Seems that LunaAppMgr does a call to mountcfs which we don't have. Seems it only exists on legacy. Sep 20 13:06:39 I have quite no idea Sep 20 13:07:04 I think it's not used on our side Sep 20 13:07:16 but not sure Sep 20 13:07:19 Let me see if they ever open sourced that Sep 20 13:07:38 If not I suggest to simply remove the call from LunaAppMgr Sep 20 13:09:52 I guess anyway we don't need it currently since it was only used in legacy to write installed apps to prevent piracy etc Sep 20 13:10:00 Not really a concern for us currently Sep 20 13:10:30 It wasn't out of concern for security ? just piracy? Sep 20 13:10:45 Well security as well I guess Sep 20 13:10:50 It was encrypted Sep 20 13:11:50 it could be useful to still have a separate mount, though Sep 20 13:12:28 If everything is already in place, I mean, why not Sep 20 13:12:30 Tofe: It seems that we have the 3.0.5 source code + palm patches Sep 20 13:12:43 We'd need to add that to the build then I guess Sep 20 13:12:51 And see if it still works :P Sep 20 13:13:08 hehe Sep 20 13:14:30 init.sh says "setup cryptofs which is not a real cryptofs yet" :P Sep 20 13:14:30 LOL Sep 20 13:14:47 I cannot find cryptofs bits in image either Sep 20 13:16:17 :D Sep 20 13:17:43 Not sure if linux provides standard functionality for an encrypted mount nowadays Sep 20 13:18:10 Last activity in http://github.com/reboot/cryptofs seems 2013 Sep 20 13:22:45 morphis: ^ You have any ideas/suggestions re cryptofs? Sep 20 13:36:28 Tofe: Well good thing is that if we go cryptofs route that there's little changes between 0.60 we're using and latest commit :P Sep 20 13:53:50 I mean 0.60 which is on legacy 3.0.5 ;) Sep 20 13:54:38 yep Sep 20 13:55:37 Just it seems that mountcfs is not open sourced, just the cryptofs stuff itself. Sep 20 13:55:55 Shouldn't be too hard to replicate it's functionality though I guess Sep 20 16:25:04 DougReeder: The Just Type field issue should be sorted :) Sep 20 16:25:41 Seems morphis changed css sometime back (probably for a good reason) but that broke it now. I changed it back since it seems to work OK with original value. Sep 20 16:36:03 Herrie|Pre3|2: outdated pong Sep 20 16:45:03 nizovn: Was able to figure out some stuff already, turned out we had some invalid JSONs and systemd order was also causing issues, that should now be history hopefully Sep 20 16:45:09 At least mainly for startup Sep 20 16:45:19 We have a few small things to tackle still but nothing too major it seems Sep 20 16:46:31 Herrie|2: if you have image with your changes, could you test luna-send -n 1 palm://com.palm.service.calendar.reminders/onInit '{}' to make sure calendar.remiders error is gone? Sep 20 16:56:54 nizovn: I still saw one error for the calendar Sep 20 16:57:02 New images are building now Sep 20 16:57:23 Mako is done, let me download & flash Sep 20 17:44:08 Hmmz I still have: https://bpaste.net/show/2703131a3277 Sep 20 17:44:51 nizovn: I guess we should fix this one too: Sep 20 08:19:19 mako LunaSysService[752]: [] [pmlog] LunaSysService JSONUTILS#156 {"FUNC":"bool LSMessageJsonParser::parse(const char*, LSHandle*, ESchemaErrorOptions)"} [Schema Error] : [msmAvail :Category: /storaged Method: MSMAvail]: Could not validate json message '{"mode-avail":true}' sent by '/var/run/luna-service2/com.palm.private_hub' Sep 20 17:44:51 against schema '{"additionalProperties":false,"properties":{"$activity":{"type":"object"},"mode-avail":{"type":"boolean"},"returnValue":{"type":"boolean"}},"required":["returnValue"],"type":"object"}'. Sep 20 17:52:41 The error i have for reminders is not on the init it seems though but earlier Sep 20 17:52:49 Could be we need to tweak some systemd as well for this Sep 20 17:57:44 ok, maybe it's something else, but good you solved some issues Sep 20 17:58:03 yes, we can fix schema in storaged Sep 20 18:04:01 nizovn: Yeah lots of issues are sorted in the log on first boot now :) Sep 20 18:04:06 All kinds are created at the right time Sep 20 18:04:12 So mediaindexer behaves as well now :D Sep 20 18:04:58 good :) Sep 20 18:05:44 I added a lot of dirs to app-services so we have a lot less log-spam Sep 20 18:19:37 nizovn: For calendar I might have an idea Sep 20 18:20:04 It seems that we don't have luna-webappmgr available when calendar is being tried to run initially Sep 20 18:20:18 So we'd need to update systemd for the right order Sep 20 18:21:39 lunappmgr should be started after lunawebappmgr it seems Sep 20 18:21:47 Lunaappmgr is launching the calendar Sep 20 18:22:08 so onInit doesn't give error after 5sec? Sep 20 18:22:51 nizovn I don't think so Sep 20 18:24:18 This is what I see on first boot: https://bpaste.net/show/4a2c665c11df Sep 20 18:25:23 yes, seems you are right about this issue Sep 20 18:26:35 Herrie|Laptop: btw could you confirm bug that Just Type can be opened on locked screen? Sep 20 18:26:41 I'll test some systemd changes for this to see if I can solve it Sep 20 18:26:48 nizovn; Yeah I also see the launchbar on bottom Sep 20 18:27:39 ok Sep 20 18:32:32 Could be simple z-index issue Sep 20 19:14:05 launchbar can be seen on lockscreen? that's not good :) Sep 20 19:14:14 Tofe: Yeah.... Sep 20 19:14:18 Not sure what causes that Sep 20 19:14:28 The first boot log is a lot cleaner BTW :P Sep 20 19:14:46 Herrie|Laptop: that's really great :) Sep 20 19:14:59 I was looking at the HTML5 File API issue Sep 20 19:15:05 quite puzzling Sep 20 19:16:25 Tofe: I added some debugging, all the Settings for the webengineview are OK Sep 20 19:16:52 I suspect it might be a path or sandbox issue Sep 20 19:17:30 But my knowledge of the depths of QtWebEngine/Blink didn't allow me to dig much further Sep 20 19:19:33 Tofe: You can use Testr in Browser app as well for the same Sep 20 19:19:55 yes, I did that, it's easy to reproduce now Sep 20 19:20:12 and I investigated inside webengine's code, but it looks quite ok Sep 20 19:20:21 local file access is enabled by default Sep 20 19:20:35 and we set it (again) to true Sep 20 19:21:35 Herrie|Laptop: you say it works in the browser, or it works in browser _and_ on the website ? (so, different origin url) Sep 20 19:21:59 I didn't test the "file from file" access witrhin the browser Sep 20 19:22:17 Tofe: http://webos-ports.github.io/org.webosports.app.testr/debug.html Sep 20 19:22:25 THat's what I tested in Browser app Sep 20 19:22:37 ok, so, from http uri ;) Sep 20 19:22:58 Tofe: Not sure what it would do if you add this to an appinfo.json :P Sep 20 19:23:46 ah, that's a good way to confirm the issue is related to 'file access from a file URL" Sep 20 19:23:48 I mean something like https://github.com/webOS-ports/web-apps/blob/master/com.twitter.web/appinfo.json Sep 20 19:23:55 And just put this URL in the main Sep 20 19:23:59 Id idn't test that Sep 20 19:24:03 Just thought about it now Sep 20 19:25:50 it works Sep 20 19:26:23 I guess the setting for AllowFileAccessFromLocalFile doesn't work well Sep 20 19:26:45 I feel an upstream PR coming :P Sep 20 19:26:51 hehe :) Sep 20 19:27:36 I'll have to debug that the hard way, I think; by simply looking at the code it *looks like* it's fine Sep 20 19:27:42 but there must be a catch Sep 20 19:29:18 Tofe few commits from upstream: https://github.com/qt/qtwebengine/commit/f589b8b411f1049f6f2f701478e0c38c72ad824a Sep 20 19:30:01 That might be worth checking: https://github.com/qt/qtwebengine/commit/5f18e75300bfeffca85c0891601fb213c4197501 Sep 20 19:30:23 2nd one might be something for our issue with the input when we have multiple tabs Sep 20 19:30:31 the first one should occur for luneos, hopefully Sep 20 19:30:38 shouldn't Sep 20 19:31:21 https://github.com/qt/qtwebengine/commit/f84869e9d8ea3ea9ced15adcc092558555874606 Sep 20 19:31:33 https://github.com/qt/qtwebengine/commit/f0a77c0f2a82147ba28eed35fcf4d1186d0aaa92 Sep 20 19:31:36 yes we need that 2nd commit Sep 20 19:32:35 3rd one I.. don't know :) Sep 20 19:32:42 last one is just c++ cosmetics Sep 20 19:34:52 I didn't check if we have them, just going quickly through dev commit history and pasting them here Sep 20 19:35:45 I guess these 3 might be good to check: https://github.com/qt/qtwebengine/commit/4c305cba0bf7cc021fd355af574b431f0d5a057d & https://github.com/qt/qtwebengine/commit/292f573f4e86d938a3bde80627637d245a949e56 & https://github.com/qt/qtwebengine/commit/aeea8816a7228e9e89fff15b26005eaaa38fce2d Sep 20 19:41:33 https://github.com/qt/qtwebengine/commit/7a49313c84ccd4779e396b7bc8341a331d90f96f & https://github.com/qt/qtwebengine/commit/d1168fda00c6eaff0f367d969639eec9adadb907 & Sep 20 19:41:40 That's all I can find for now Sep 20 19:41:45 Could be we have some already Sep 20 19:52:54 activeFocusOnPress is always true for us Sep 20 19:53:53 last one is the same as first one, it seems :) Sep 20 19:54:11 sandboxing on Linux would be interesting, but I have no idea what the consequences might be **** ENDING LOGGING AT Wed Sep 21 02:59:58 2016