**** BEGIN LOGGING AT Mon Apr 04 02:59:58 2016
Apr 04 04:53:20 <Herrie> morning
Apr 04 07:03:21 <Herrie|Veer> morning
Apr 04 07:41:39 <Tofe> Morning!
Apr 04 07:44:52 <Herrie|Veer> Tofe: ICU error is gone now :)
Apr 04 07:45:07 <Herrie|Veer> With latest build :)
Apr 04 07:50:58 <Tofe> ok, that's good
Apr 04 07:51:06 <Tofe> so we switch branches?
Apr 04 07:51:29 <Tofe> -next => trash, tofe/work => -next ?
Apr 04 07:52:10 <Herrie|Veer> Still crashes on html5test.com with ARM targets though. qemux86 works fine. I suspect libc 2.22.
Apr 04 07:52:16 <Herrie|Veer> Tofe: Yeah I guess so.
Apr 04 07:52:30 <Tofe> Herrie|Veer: maybe you could help me with a choice for a LuneOS build harddrive
Apr 04 07:53:06 <Herrie|Veer> Tofe: Really depends. I would get a SSD ;)
Apr 04 07:53:09 <Tofe> I hesitate between a small SDD (120Gb) and a confortable HDD (2Tb)
Apr 04 07:53:34 <Herrie|Veer> I didn't have good experience with Samsung 850 in my config. I have Sandisk and seems decent.
Apr 04 07:53:40 <Herrie|Veer> A lot more stable.
Apr 04 07:55:12 <Tofe> The thing is, I'd like to keep sstate and work directories as long as possible, as it shortens the build a lot
Apr 04 07:55:30 <Tofe> It won't be possible with a small SSD
Apr 04 07:56:10 <Herrie|Veer> I got a Ultra II 480GB one from Sandisk, paid EUR 120,90 for it a while back when they had a special deal on Amazon.fr
Apr 04 07:56:19 <Herrie|Veer> It's around 137 now it seems.
Apr 04 07:56:28 <Tofe> that's a quite good price you got here :)
Apr 04 07:57:03 <Herrie|Veer> Yeah :)
Apr 04 07:57:20 <Herrie|Veer> It's not the fastest SSD out there but it's reliable at least :)
Apr 04 07:57:27 <Herrie|Veer> 120GB was really too small for me.
Apr 04 07:57:51 <Herrie|Veer> And it's still a lot quicker compared to a 7200 or 10k rpm regular hdd :P
Apr 04 07:58:05 <Tofe> For a build drive I don't think the top bandwith is important; it's more about the access time, which is already quite low even on low-end devices
Apr 04 07:59:11 <Herrie|Veer> I have it for my Windows 10 and have 200GB allocated in VirtualBox for my Ubuntu 15.10 build :)
Apr 04 08:01:20 <Tofe> http://www.amazon.de/SanDisk-Plus-480GB-Interne-schwarz/dp/B018X8LB2G/ref=sr_1_3?s=computers&ie=UTF8&qid=1459756852&sr=1-3 ?
Apr 04 08:02:40 <Herrie|Veer> Ultra II is 126,90 on Amazon.de :)
Apr 04 08:02:55 <Herrie|Veer> You should check performance difference between the 2 I guess
Apr 04 08:03:06 <Tofe> yes, I don't know yet what's the difference
Apr 04 08:04:54 <Herrie|Veer> http://storagereview.com is a good site for tests :)
Apr 04 08:05:28 <Herrie|Veer> Seems Plus = entry level. Ultra II above that.
Apr 04 08:05:44 <Herrie|Veer> Extreme is top of the line, but also more expensive.
Apr 04 08:10:01 <Tofe> Well, frankly, it's ~450MB/s vs ~550MB/s
Apr 04 08:10:24 <Herrie|Veer> Well and mainly write speed difference it seems
Apr 04 08:10:53 <Herrie|Veer> IOPS is higher on Ultra II
Apr 04 08:11:14 <Herrie|Veer> And 5 vs 3 year warranty it seems ;)
Apr 04 08:11:14 <Tofe> Ah, that could be a good reason to buy UltraII
Apr 04 08:11:29 <Tofe> (the IOPS, not guarantee)
Apr 04 08:11:36 <Herrie|Veer> Not that I ever needed it on a SSD so far :)
Apr 04 08:11:50 <Herrie|Veer> But was happy to have 5 yrs on some of my regular HD's :)
Apr 04 08:13:14 <Herrie|Veer> I went from Sandisk 120GB->Samsung (due to good reviews)->Sandisk. I had too many stability issues with Samsung in my config.
Apr 04 08:13:57 <Herrie|Veer> Especially with my raid array. Then again could be just my setup. But my experience with Sandisk has been good, also for SD cards etc
Apr 04 08:15:26 <Herrie|Veer> Knowing Samsung & software (*TouchWiz*) I blame it on their software :P
Apr 04 08:17:48 <Tofe> :)
Apr 04 08:23:25 <Herrie|Veer> You never have enough storage and 480GB will take you a lot longer to fill compared to 120GB. Prices are reasonable nowadays as well :)
Apr 04 08:30:55 <Tofe> that's true
Apr 04 08:48:47 <Tofe> https://together.jolla.com/question/132592/patch-for-miniaturized-views-of-actual-app-instead-of-covers/ hehe!
Apr 04 08:49:45 <Herrie|Veer> 120GB in my vbox filled up quickly. 200 is no luxury so I guess you'll be happy with 480 :)
Apr 04 08:50:27 <Tofe> yes, I've already taken the decision :) I'll just wait a couple of days to synchronize the delivery with the rest  of the upgrade :p
Apr 04 08:51:03 <Tofe> Otherwise I'll be really tempted to launch a build, which will make the cpu go high on temperature :p
Apr 04 08:57:51 <Herrie|Veer> LOL :P
Apr 04 08:58:15 <Herrie|Veer> Amazon is quick in general :) I get things from Germany in 2 days usually :)
Apr 04 09:21:08 <Herrie|Veer> JaMa: ping
Apr 04 10:14:39 <JaMa> Herrie|Veer: pong
Apr 04 10:20:24 <Herrie|Veer> JaMa: 3 things :P
Apr 04 10:20:49 <Herrie|Veer> 1. Seems tenderloin build is failing. Seems it cannot fetch something from sstate :s
Apr 04 10:21:34 <Herrie|Veer> 2. Migration to QT 5.6 & Jethro is completed (Just it's still 5.6 RC and not final). Any ETA on final?
Apr 04 10:24:06 <Herrie|Veer> 3. Seems we get funny segfaults on various components (media indexer, ipkg-service). References are to libc-2.22.so. We also have hard crashes of qtwebengine on ARM targets for html5test.com but it works ok for qemux86. The segfaults for components happen
Apr 04 10:24:06 <Herrie|Veer>  on both qemux86 and ARM
Apr 04 10:51:20 <JaMa> 1. the errors aren't important or fatal for image creation NOTE: Tasks Summary: Attempted 6454 tasks of which 6161 didn't need to be rerun and all succeeded.
Apr 04 10:52:30 <JaMa> 2. I know, not sure when I'll update meta-qt5, there are only few commits between SRCREV there and final 5.6 in same cases it's actually newer revision (from 5.6 branch than what's 5.6.0)
Apr 04 10:53:02 <JaMa> 3. do you have example backtrace? Why are they funny?
Apr 04 12:03:57 <Herrie|Veer> JaMa: We bumped webengine and chromium in our fork to latest 5.6 from upstream.
Apr 04 12:04:07 <Herrie|Veer> So it's more about the other QT bits.
Apr 04 12:04:27 <Herrie|Veer> So 5.6.0 = final 5.6 release and 5.6 is WIP for 5.6.1 etc?
Apr 04 12:07:58 <JaMa> Herrie|Veer: yes, 5.6.1 will be branched from 5.6 branch when time comes
Apr 04 12:08:14 <JaMa> as 5.6.0 was earlier
Apr 04 12:10:25 <Herrie|Veer> I'm not that familiar with c(++) and gdb, if you have instructions I'm happy to provide a backtrace.
Apr 04 12:10:44 <Herrie|Veer> Ah seems they didn't create a 5.6.0 branch, just tagged the release.
Apr 04 15:12:17 <Herrie|Veer> nizovn: ping
Apr 04 15:12:42 <nizovn> Herrie|Veer: pong
Apr 04 15:13:56 <Herrie|Veer> Do you have chance to run some backtraces on latest virtualbox image? Seems mediaindexer & ipkgservice are segfaulting due to libc-2.22.so somehow :s
Apr 04 15:14:14 <Herrie|Veer> I'm a novice with gdb, so I'm always lost :P
Apr 04 15:15:00 <nizovn> yes, i can test
Apr 04 15:15:40 <Herrie|Veer> mediainder is libglib-2.0.so actually.
Apr 04 15:16:26 <Herrie|Veer> ipkgservice is libc-2.22.so. Seems it has to do with Fido-> Jethro upgrade.
Apr 04 15:16:50 <Herrie|Veer> It's mostly stable we just have a few of these. For Preware2 just install LuneTube to get it :)
Apr 04 15:18:20 <nizovn> ok, will see
Apr 04 15:19:08 <Herrie|Veer> If you can paste the results somewhere JaMa can have a look. It might just ring a bell for him ;)
Apr 04 15:32:47 <nizovn> mediaindexer crash seems only to appear on first boot. could be easily because of not initialized database somehow(i think).
Apr 04 15:33:14 <Herrie|Veer> nizovn: Yeah that could be. First boot needs some tweaking.
Apr 04 15:33:39 <Herrie|Veer> For Preware2. You might need to rotate with F7 to get list of applications.
Apr 04 15:34:32 <Herrie|Veer> nizovn: segfaulting still isn't very graceful right :P ?
Apr 04 15:36:05 <nizovn> yes..
Apr 04 15:36:48 <JaMa> Herrie|Veer: should I fix "testing" branch in webos-ports-setup?
Apr 04 15:37:49 <Herrie|Veer> JaMa: Fix what exactly?
Apr 04 15:40:07 <JaMa> to be based on "jethro" branch
Apr 04 15:40:28 <Herrie|Veer> JaMa: I did already ;)
Apr 04 15:40:33 <JaMa> you were supposed to cherry-pick rpi support from fido to jethro branch
Apr 04 15:40:39 <JaMa> then rebase testing branch on top of jethro
Apr 04 15:40:48 <JaMa> that's not what "testing" is now
Apr 04 15:41:23 <Herrie|Veer> I only see jethro references? qt5 is pointing to master for qt 5.6
Apr 04 15:42:02 <Herrie|Veer> Glad to learn what I've done wrong :P
Apr 04 15:42:37 <nizovn> this is the backtrace of ipkservice crash https://bpaste.net/show/e9ffad2ded50 . i was bored hitting enter, so it's not complete :)
Apr 04 15:43:14 <JaMa> Herrie|Veer: you're missing other changes I had prepared in jethro branch of webos-ports-setup
Apr 04 15:43:34 <JaMa> like tag name in scripts/staging_new.sh, removal of unused ruby layer
Apr 04 15:44:04 <Herrie|Veer> JaMa: Ah, yeah please fix those :)
Apr 04 15:45:34 <Herrie|Veer> We've PR-ed the single commit we have in meta-raspberrypi to upstream. Once they merge we can use that directly.
Apr 04 15:47:16 <Herrie|Veer> nizovn: Thnx
Apr 04 15:47:23 <Herrie|Veer> JaMa: ^ Of any use?
Apr 04 15:48:39 <JaMa> also https://github.com/webOS-ports/meta-raspberrypi was pointing to master branch
Apr 04 15:49:39 <JaMa> nizovn: without dbg symbols this bt doesn't say much
Apr 04 15:49:44 <JaMa> but it's way too long
Apr 04 15:51:14 <Herrie|Veer> JaMa: As per Andolamin we need master since jethro misses a lot.
Apr 04 15:51:22 <Herrie|Veer> Tested & confirmed to be working
Apr 04 15:53:38 <JaMa> why not merge required changes to jethro branch?
Apr 04 15:53:49 <JaMa> or are these changes needed only for meta-qt5/master?
Apr 04 15:54:30 <JaMa> ah because you want to use upstream meta-raspberrypi layer?
Apr 04 15:55:52 <Herrie|Veer|2> JaMa: Yes because upstream doesn't maintain Jethro anymore.
Apr 04 16:02:27 <JaMa> updated testing branch pushed, I've switched unstable to point to master builds
Apr 04 16:03:13 <Herrie|Veer|2> JaMa: Thnx
Apr 04 17:36:40 <nizovn> Herrie|2: the issue with ipkgservice was caused by too big string reply from curl(?). it was 1281 chars long, so i changed https://github.com/webOS-ports/preware/blob/master/oe-service/src/luna_methods.h#L32 to 4096, and it worked.
Apr 04 17:36:51 <nizovn> not sure it's the best solution though
Apr 04 17:38:22 <Herrie|2> nizovn: Well that should work for a while?
Apr 04 17:40:53 <Herrie|2> nizovn: Strange that broke frm Fido to Jethro though
Apr 04 17:44:24 <nizovn> yes, probably
Apr 04 18:44:29 <Herrie|2> Tofe/JaMa: Seems RPi guys are happy to accept our PR. I just needed to send it to the mailing list, which I just did.
Apr 04 18:53:35 <Tofe> Herrie|2: great!
Apr 04 18:54:18 <Tofe> Herrie|2: "Jethro" is 2 more characters than "Fido" ?... Ok, that's a long shot :)
Apr 04 19:08:34 <JaMa> Herrie|2: you commit on ML looks broken by your mail client, try to use git send-email
Apr 04 19:09:06 <JaMa> otherwise most layer maintainers will just refuse such change
Apr 04 19:09:11 <JaMa> I would :)
Apr 04 19:45:23 <Herrie|2> Tofe: You OK with the change nizovn proposed?
Apr 04 21:52:57 <elvispre> Andolamin: Thanks for your Enyo 2.7 assistance yesterday. I am now up and running with a converted app (in the browser, at least).
Apr 04 21:53:57 <elvispre> Andolamin: Your 'cast' app has appinfo.json, and icon*.png as assets in the top level folder. Is that a webOS / luneOS  thing?
Apr 04 21:55:50 <elvispre> Andolamin: BTW, I have a webOS TV. Am I interested in what cast does? :-)
Apr 04 21:58:22 <Andolamin> elvispre: Not sure why I put icons and appinfo at top level, always have. Must have picked that up ~5 years ago. com.alanstice.cast will eventually go into my Preware feed to act as a webOS/Chromecast second screen service for LuneOS
Apr 04 23:18:58 <EricBlade> isn't appinfo.json supposed to be in the root directory of the app? that's where i always put it
Apr 04 23:38:23 <Andolamin> DougReeder: Just wanted to let you know I haven't forgotten about your email. I am waiting for approval from my boss so I'm not making any statements that aren't backed up.
**** ENDING LOGGING AT Tue Apr 05 02:59:58 2016