**** BEGIN LOGGING AT Tue Oct 11 02:59:58 2016 Oct 11 03:13:38 morning Oct 11 03:14:00 DougReeder: Currently the RPi3 runs in 32 bit only, hence compatibility with the 2 Oct 11 03:14:39 The nightly builds currently don't build for the pi Oct 11 03:14:48 Andolamin is looking into that Oct 11 03:16:34 * DougReeder nods Oct 11 03:17:05 You’re up early. :-) Oct 11 03:18:24 Someone decided 4:30 is a nice time... We managed to get him down for another 30 or so mins, but that was it :P Oct 11 03:18:38 :-( Oct 11 03:21:34 Ours insisted on sleeping the playfort. Hey, as long as sleep happens! Oct 11 03:24:56 And he went to bed after 21:00 :S Oct 11 03:40:54 Any suggestions on the gstreamer - valgrind problem? Oct 11 03:43:13 This is for the Pi? Oct 11 03:43:22 No, for N4. Oct 11 03:43:28 DougReeder: Hmmz Oct 11 03:43:32 That should work normally Oct 11 03:43:37 :-S Oct 11 03:45:32 I guess you tried to re-run it? Oct 11 03:45:52 * DougReeder nods “Several times, alternating with make update” Oct 11 03:52:02 Hmmz Oct 11 03:52:12 Let me try at my end Oct 11 03:52:21 Thank you. Oct 11 04:02:01 Did you try DEVICE=mako bb -f gstreamer1.0-libav -c cleanall Oct 11 04:02:13 And then: DEVICE=mako bb -f gstreamer1.0-libav Oct 11 04:02:17 Nope. I’ll try that Oct 11 04:02:22 That might work sometimes Oct 11 04:03:55 First runs without errors Oct 11 04:04:03 It should Oct 11 04:04:04 Second warns me: Oct 11 04:04:11 WARNING: /home/doug/testing/webos-ports-env/webos-ports/openembedded-core/meta/recipes-multimedia/gstreamer/gstreamer1.0-libav_1.6.3.bb.do_build is tainted from a forced run Oct 11 04:04:15 Ah that's OK Oct 11 04:04:20 :-S Oct 11 04:04:25 THat's because you force it to compile manually Oct 11 04:07:28 Tofe: The spreading works a lot better now Oct 11 04:08:58 Tofe: Just it seems that when you have a stack of 3, the 2nd card should be in the middle, it now seems to use the 1st card as an anchor, causing the 3rd card to go off screen at the right side Oct 11 04:09:10 Tofe: When you spread Oct 11 04:16:23 That seems to have unblocked the build; thanks! Oct 11 04:16:49 DougReeder: Yeah it sometimes happens.... This was a completely clean build? Oct 11 04:16:55 Or an update of a previous one? Oct 11 04:17:02 No, an update. Oct 11 04:17:09 We've been able to tackle a lot of build dependency issues in the past already Oct 11 04:17:36 Because before it would largely depend on the number of concurrent tasks that were running if you would run into dependency issues Oct 11 04:17:37 I’m all in favor of 1-button builds. Oct 11 04:18:02 I did a huge cleanup earlier this year to solve most of those Oct 11 04:18:24 So am I, especially for users who try stuff for the first time it should just work :P Oct 11 04:19:56 * DougReeder nods. “Not fixing that is a big frustration for me at the place I currently work” Oct 11 04:27:30 Heh, I reported a bug against the Alarm API in Firefox OS in 2014, and it’s being closed now, because the API is being removed. :-S Oct 11 04:28:11 DougReeder: Hmm :'( Oct 11 04:28:34 C’est la vie. Oct 11 04:32:58 Not too pleased with Mozilla's decisions lately Oct 11 04:33:04 Thunderbird, FirefoxOS etc Oct 11 04:33:31 They're even considering using other rendering engines which might mean they'll drop their own :S Oct 11 04:36:13 That would be huge - but they have to prioritize, like everyone else. Oct 11 04:36:58 Yeah I understand that, but they seem to drop the ball on a lot of things lately Oct 11 04:37:59 One can’t fault the quantity of effort they put into FFOS, but it could have been better directed. Oct 11 04:38:46 They seemed to forget the part about offering a better user experience. Oct 11 04:39:55 With Progressive Web Apps, now would be a better time for FFOS than a couple years ago. Oct 11 04:41:12 And little of the work I did on an app for FFOS (Sernene Notes) got wasted. I can’t say that about webOS. Oct 11 04:49:58 Yeah :S Oct 11 05:05:17 Well, I guess I won’t be commenting on the graphical improvements PR Oct 11 05:08:48 DougReeder: It fixed a lot of small issues ;) Oct 11 05:08:53 Nothing too major really Oct 11 05:09:19 I could only have commented on the visuals, anyway. Oct 11 05:09:45 DougReeder: The red links I don't really like, but well ;) Oct 11 05:09:56 Can always find a more suitable color Oct 11 05:10:20 In web pages? Oct 11 05:10:50 No just on the license page in first use for links Oct 11 05:12:42 Toggle switches are still really hard to turn on. Oct 11 05:13:35 The links show blue for me. Oct 11 05:13:45 DougReeder: I compared previous with current and I really didn't find difference Oct 11 05:13:54 When you reported it last time Oct 11 05:14:04 Hmm Oct 11 05:14:05 DougReeder: New PR for FirstUse isn't in image yet ;) Oct 11 05:14:29 Ah, of course; I ran make update an hour or two ago. Oct 11 05:14:45 Yeah Oct 11 05:28:29 Hrm, it’s not clear the the country defaults to the US, now Oct 11 05:28:47 And the time zone definitly doesn’t default to the New York. Oct 11 05:29:12 Or is it just that the lists don’t scroll to those locations? Oct 11 05:30:01 And my N4 is not finding any WiFi networks to connect to. Oct 11 05:30:32 Sorry to report problems and leave, but I’m up past my bedtime. Oct 11 05:32:22 OK I'll double check Oct 11 05:47:27 Yeah you seem to be right, I'll debug that a bit Oct 11 07:55:14 Morning! Oct 11 07:56:47 Herrie|Pre3: ah yes, good remark, it might be that the spreading isn't done on both sides Oct 11 07:57:08 or, more precisely, that the horizontal centering isn't done properly Oct 11 07:57:30 Tofe: Yeah it seems that way ;) Oct 11 07:58:00 Herrie|Pre3: is your PR already tested? Oct 11 07:59:23 Tofe: Which one? Luna-sysmgr? Oct 11 07:59:25 Yeah :) Oct 11 07:59:31 Stuff still boots :P Oct 11 08:08:35 :) Oct 11 08:10:07 2 more to go ;) Oct 11 08:10:25 Then webos-systemd-services is empty :P Oct 11 08:15:04 Tofe: Can you check PR 198 in meta-webos-ports? Oct 11 08:15:29 I confirm it's working, just if you can check the luna-sysmgr changes are OK with you Oct 11 09:41:43 Herrie|Pre3: For N5, does this article mean what I think it means ? http://www.phoronix.com/scan.php?page=news_item&px=Linux-4.9-ARM-Pull Oct 11 09:42:57 ah no wait, we would still have issues with drivers. Oct 11 09:44:26 Tofe: Yeah but it should help when we can use a mainline kernel I guess Oct 11 09:44:37 Should make some things easier at least? Oct 11 10:42:17 Not sure we'll even be able to use it, because of the android drivers Oct 11 11:48:48 Tofe: It really depends. The RaspberryPi's also don't use Android drivers. Really depends what's available for the N5. Oct 11 13:00:29 DougReederPhone: ping Oct 11 14:50:50 Herrie|Pre3: for the calendar icon PR in luna-next, I'm not yet sure: should we hack the compositor to have proper calendar icons ? What about the full launcher icons ? Can't we get that dynamic behavior somehow differently ? Oct 11 14:52:07 I'll try to find how legacy did it; it could be it was hard-coded indeed. Oct 11 15:16:18 Tofe: I can look it up but it was nastier with even having an activity for it! Oct 11 15:20:08 I'll find it for you Oct 11 15:39:53 Tofe: com.palm.app.calendar.firstrun.json: "sets calendar icon to have the current date" :P Oct 11 15:48:58 Tofe: I'll send you a gem of code :P Oct 11 15:55:42 I guess we could use that as well... It seems to call enyo.palmServices.application with a method called updateLaunchPointIcon Oct 11 15:57:28 Let me see where that hides and if we have it. Oct 11 16:12:17 Tofe: Seems we can do that with a call to applicationManager. Just not sure how we'd do this in LNC. I guess we'd call applicationManager and get the icon from there. Oct 11 16:12:21 I'll rework it a bit. Oct 11 16:12:29 Shouldn't be rocket science. Oct 11 16:29:44 morning Oct 11 16:51:14 nizovn: morning Oct 11 16:51:24 I erased your FirstUse commit from history Oct 11 16:51:32 Had quite some issues we found out after merge Oct 11 16:51:43 I did a bit of digging but seems it needs some rework Oct 11 16:51:50 So you can update & PR again Oct 11 16:51:54 WiFi scan doesn't work for me Oct 11 16:51:58 That's the biggest Oct 11 16:52:47 Then we have https://github.com/webOS-ports/luna-init/blob/webOS-ports/master/files/conf/locale.txt#L109 and https://github.com/webOS-ports/luna-init/blob/webOS-ports/master/files/conf/region.json#L137 not being equal Oct 11 16:53:25 I need to cross reference the 2 JSON files to make sure that all country names are equal. We could do something in the LNC code, but that would always be a nasty workaround Oct 11 16:54:58 but wifi scan works if revert PR? Oct 11 16:55:04 Then we have things like https://github.com/webOS-ports/luna-init/blob/webOS-ports/master/files/conf/locale.txt#L506 that cause issues Oct 11 16:55:13 But probably always have anyway Oct 11 16:56:17 i'm not sure how these files are used in firstuse.. Oct 11 17:00:37 Herrie: thanks for the digging Oct 11 17:00:57 what I'd like to push is a way that could be implemented by any other app, potentially Oct 11 17:03:01 nizovn: I didn't test wifi but it worked with previous nightly so I have no reason to assume it won't now. Oct 11 17:03:20 These files feed the getPreferenceValues initially. Oct 11 17:03:43 So as long as we make sure they're aligned between the 2 we should be fine. Oct 11 17:04:08 We didn't touch the wifi code for a while. Oct 11 17:05:22 Just a few of the languages with cyrl and latn will have issues. We don't really have a solution for that now I think implementation wise. I.e. Serbian can have Latin and Cyrillic kb layout, same for Uzbek and a few others it seems. Oct 11 17:06:20 I guess we could go for the most common option for now and only have 1 for each language and put a feature request in the issue tracker for the future to properly address it when needed. Oct 11 17:06:32 It's anyway for quite obscure languages :P Oct 11 17:07:11 Tofe: OK, out for dinner with colleagues tonight but might get to it later or tomorrow morning. Generic approach shouldn't be too hard to implement I guess. Oct 11 17:09:46 I hope so, I'll have a look too on QML side Oct 11 17:10:51 Wifi scan woked fine for me, with code from an hour or two before the PR Oct 11 17:11:15 ok Oct 11 17:11:27 Tofe: I guess we could do a 1 time query for all appInfo info store that in a local array and then fetch the icon for each app from this array Oct 11 17:11:46 Instead of doing multiple ls calls Oct 11 17:11:58 We do that for the full launcher icons Oct 11 17:12:02 iirc Oct 11 17:12:15 so the array isn't very far away Oct 11 17:13:09 Tofe: Yeah, should probably be a few lines of code. Just wanted to make sure we have the "right" apps as well, so took those from legacy 2.x and 3.x Oct 11 17:14:04 We actually have an error/warning in journalctl about it not being able to find an activity therefore cannot cancel/update/replace it. Might actually be related to this calendar stuff :P Oct 11 17:14:08 Will dig a bit ;) Oct 11 17:14:40 So the code exploring might have been useful in multiple areas :P Oct 11 17:15:41 Tofe: I'm also able to get Out Of Memory on first boot after First Use and going to Accounts app and adding my gmail account. Seems we might be leaking somewhere :P Oct 11 17:15:47 On 2nd boot it behaves. Oct 11 17:15:59 mmh, ok, interesting Oct 11 17:16:07 Can I use top to see what's eating memory or you have other suggestions? Oct 11 17:16:20 I was able to replicate this a number of times on N4 Oct 11 17:16:24 Pretty consistently. Oct 11 17:16:24 top should be fine for leak that large Oct 11 17:16:32 Tofe: OK Oct 11 17:16:42 Will have a look probably tomorrow morning Oct 11 17:16:49 ok Oct 11 17:18:14 There are still a few minor systemd startup tweaks to do it seems, but I think we've solved 90-95% of the issues now :) Oct 11 17:18:21 Next is to speed up the shutdown a bit. Oct 11 17:18:30 And maybe tweak the startup speed a bit. Oct 11 17:18:59 Errors in the log are mainly gone now it seems :) Oct 11 17:20:27 Herrie|Pre3: seems i will need some help with testing wifi stuff on n4, as i can't replicate this on TP Oct 11 17:20:42 nizovn: OK Oct 11 17:21:01 It should behave the same.... Oct 11 17:21:03 so you flashed new image, and wifi page was never updating? Oct 11 17:21:10 nizovn: Yeah Oct 11 17:21:25 Well I just pushed you FirstUse code Oct 11 17:21:34 But DougReeder tested a whole image Oct 11 17:21:41 I did afterwards as well. Oct 11 17:22:18 I did have to force clean and build one component Oct 11 17:22:51 so how firstuse app was launched? Oct 11 17:22:52 But that was gstreamer Oct 11 17:23:19 nizovn: I tried clean flash as well, same issue Oct 11 17:23:26 After DougReeder reported it Oct 11 17:23:28 ok Oct 11 17:23:42 First time I used the com.palm.applicationManager/launch Oct 11 17:29:35 Herrie|Pre3: could you please see in logs this message https://github.com/nizovn/org.webosports.app.firstuse/blob/a2c8b8ad84fa1e014588c619771f808b8e34876d/qml/WiFiPage.qml#L57 ? Oct 11 17:36:58 nizovn: Will check that when home Oct 11 17:37:05 ok, thanks Oct 11 21:12:28 nizovn: https://bpaste.net/show/4ffb3919b590 Oct 11 21:12:37 This is on first boot with your FirstUse Oct 11 21:13:39 Tofe: PR looks fine on initial testing Oct 11 21:13:49 Will play a bit more in the morning before merging Oct 12 00:08:30 I just pulled the latest code and rebuilt. On my N4, First Use shows US and NY as defaults once more. Oct 12 00:09:11 WifFi shows local neworks,and the links are in blue once more. Oct 12 00:12:09 Spread to adjust card spacing didn’t work before, and it doesn’t work now. Oct 12 00:14:52 The build complained that “ Package version for package org.webosports.app.firstuse went backwards”, but that was the point, yes? **** ENDING LOGGING AT Wed Oct 12 02:59:58 2016