**** BEGIN LOGGING AT Sun Dec 28 02:59:59 2014 Dec 28 08:00:47 so.. i'm trying to decide between dedicating my TP 4G or my spare WiFi TP to Lune. Dec 28 08:01:49 whichever one i don't dedicate to Lune is getting Android 5. I don't have the 4G in service, but I could. Dec 28 08:27:13 EricBlade: Data is not 100% working for now so I'd use the Wifi one ;) Dec 28 08:38:15 last i knew it was 0% working in android 5 Dec 28 08:45:39 Ah ok... Well in LuneOS it almost works :P Dec 28 08:45:48 There's no UI yet, but could my Sierra 7710 to work ;) Dec 28 08:45:57 And confirmed Ericsson to be working as well ;) Dec 28 08:46:15 morphis is working on UI part to be integrated Dec 28 08:52:02 ok, so it really wouldn't be exceptionally helpful for you guys to have me running it Dec 28 08:54:10 Well you know a bit more kernel stuff then me :P Dec 28 08:54:20 So in that regard it would help :P Dec 28 08:54:36 And mine has a broken digitizer board :S Dec 28 08:54:53 I managed to break it while swapping the Ericsson and Sierra which I did a ton of time before already :S Dec 28 08:55:11 the last anything i did on kernel was submitting a patch back in 2002 that was necessary to get SMP running on a Compaq machine that had 2 CPUs. :-S Dec 28 08:56:48 Still searching for a replacement digitizer board... Dec 28 08:57:07 Swapped it with my wife's wifi one which had water damage, it's usable but doesn't work 100% Dec 28 08:59:59 well, i do have that one touchpad that is of questionable usability. it's almost impossible to get it to boot, and i don't think it will actually charge, so i'm not entirely certain if it's possible to doctor it Dec 28 09:00:51 but on the otherhand, even if it's not usable, it's sort of a relic of webOS, considering the webOS media system was written using it Dec 28 09:16:46 You might be able to reflash the a6 with tptoolbox? Dec 28 09:28:17 any attempt to use it will basically automatically destroy it's uniqueness. so, use it or let it be a museum piece.. hmm. Dec 28 09:39:29 Yeah that's true.. Least invasive you could do is connect it via charger, memboot it and run PmA6Updater to reflash/calibrate A6 Dec 28 09:39:48 Read reports that worked on some TP's Dec 28 09:52:09 hmm. that might get it to charge/boot consistently. oh well. right now all the chargers are charging up the two touchpads that are about to get lobotomized Dec 28 09:55:03 and then the chargers are going to get to working on some Go batteries that I have no idea if they work or not Dec 28 09:58:18 I see :) Dec 28 09:58:33 Well Go batteries usually work, they might just take quite some time to show a sign of life Dec 28 09:59:07 From the 10 or so I had here all looked completely dead... Leaving them a few hrs on the charger showed the white flashing light, then a while later the battery symbol Dec 28 10:38:37 well, sadly, the one go that i've tried so far, it didn't have a battery when i got it. the charge indicator would come on, but nothing happened. now that it has a battery, nothing happens at all. so i figure i need to at least fiddle with it some before giving up Dec 28 10:40:39 hmm. ok, these two touchpads have been charging for almost the exact same amount of time. the wifi one was at about 4% when i started. The 4G was dead. The wifi is up to 91%, the 4G is at 50% :-S Dec 28 10:42:12 Using TouchStone or barrel charger? Dec 28 10:42:26 touchstone Dec 28 10:44:20 That could explain,but also dead battery takes a lot longer to get into charging state Dec 28 10:45:31 yeah.. it was only off for maybe a minute before it came on though. but then booting probably sucked 10% lol Dec 28 10:46:15 Yeah LOL Dec 28 10:47:35 For your messaging bits, have you been rewriting SynerGV in Enyo 2? Saw you being quite active in EnyoJS channel recently :P Dec 28 10:47:46 i have Dec 28 10:48:41 http://www.ericbla.de/phpBB3/viewtopic.php?f=3&t=26#p1901 .. let me know if that link gives you a forbidden or some other error Dec 28 10:49:11 no it won't be pink. Dec 28 10:50:11 :D Dec 28 10:50:14 Nice:D Dec 28 10:50:29 Enyo 2 layout is not as nice option wise as Enyo 1 but well :P Dec 28 10:55:08 yeah, i'm not a huge Onyx fan. I was thinking of poking at Mochi, but from what I've seen from traffic here, it sounds like it needs work even to start. Dec 28 10:55:53 Well samples all work except for 1 it seems Dec 28 10:56:00 Has been updated to 2.5.1 recently as well ;) Dec 28 10:56:07 oh, is there a web sampler? Dec 28 10:56:13 Let me see Dec 28 10:56:17 Not 100% sure Dec 28 10:57:33 I want to do some experimentation with doing different UIs on different platforms but with having enough reusable stuff that i'm not completely rebuilding the entire application for each platform Dec 28 11:00:21 hoping that I can handle much of the differences between say iOS and Android with style sheet changes and some conditionally created components, but to go to something like WP or BB10 or webOS TV, that really necessarily needs vastly different UI Dec 28 11:02:13 ... this touchpad just dinged that the battery was fully charged. i don't recall webOS doing that. Dec 28 11:02:27 I can create the GH-pages branch, however it seems taht it doesn't exist for enyo and layout folders, so not very useful Dec 28 11:02:40 EricBlade: I know the phones do Dec 28 11:02:48 I dont usually pay attention to the TP's Dec 28 11:02:56 Since they charge on Touchstone mainly Dec 28 11:03:26 maybe they always did i just never knew what it was making a bing noise at me for Dec 28 11:05:36 ok, so, i guess i need to figure out how to install Lune again now Dec 28 11:05:53 http://webos-ports.org/wiki/Template:Tenderloin_Install_Steps Dec 28 11:06:22 oh thanks Dec 28 11:07:02 huh. this doesn't have preware. is it even in dev mode? heh Dec 28 11:07:32 And for Tailor you need the alpha/beta feeds enabled Dec 28 11:07:47 http://www.webos-internals.org/wiki/Testing_Feeds Dec 28 11:13:04 huh. again, first time i ran Tailor, just got a white screen, closed and restarted, and it worked. weird. Dec 28 11:18:41 1.96gb good? hope so Dec 28 11:19:12 Yeah should be fine Dec 28 11:19:16 Image is 300MB or so Dec 28 11:35:11 by the time this is done booting, the next release will be out :-D Dec 28 11:37:42 other than a lot of the graphics being surprisingly ugly, this is a huge improvement from the last build i tried. good work Dec 28 11:39:21 ... until the browser caused a reboot .. heh Dec 28 11:44:49 ok it's very crashy. Dec 28 11:47:06 TP has some memory issue it seems Dec 28 11:47:32 Could be due to 4G Kernel we're using. Morphis is looking into it Dec 28 11:47:54 Causes reboots when it runs out of memory. Runs fine on N4 though and was OK with earlier build Dec 28 11:47:59 ok.. my adb isn't liking it.. i'm not enjoying having to reinstall the android adb driver every time i plug in a new device Dec 28 11:48:01 Graphics is being worked on ;) Dec 28 11:48:51 hmm. on first use, it was able to login to my wifi, but apparently lost the setting on reboot. loading Settings from the WiFi sys menu seems to kill it Dec 28 11:50:30 whoa.. my PC has MAP bluetooth drivers now. Dec 28 11:55:24 no logs? Dec 28 12:00:14 log daemon doesn't start on boot Dec 28 12:00:24 might not even matter, if no one is using it Dec 28 12:02:00 Herrie: hellllp :D how do i get the wifi working? Dec 28 12:02:20 EricBlade: there is no log file in /var/log Dec 28 12:02:26 you have to use journalctl Dec 28 12:05:35 i'm guessing that LunaSysMgr is here to provide a bunch of bus calls that otherwise wouldn't work? Dec 28 12:07:05 EricBlade: yeah Dec 28 12:07:16 but LunaSysMgr is split into LunaSysMgr and LunaAppManager Dec 28 12:07:33 i tried killing it to see what happened. it didn't like it :-) Dec 28 12:08:00 LunaSysMgr provides things like com.palm.display Dec 28 12:08:09 and LunaAppManager mainly com.palm.applicationManager Dec 28 12:08:13 and org.webosports.bootmgr Dec 28 12:08:33 journalctl > /media/internal/mylogfile.tt Dec 28 12:08:38 journalctl > /media/internal/mylogfile.txt Dec 28 12:08:53 Then use adb pull /media/internal/mylogfile.txt Dec 28 12:09:16 ok, there's a couple of things that i feel like i really need before i can actually work on accomplishing something .. WiFi, and a terminal that doesn't suck. Dec 28 12:09:19 or use journalctl --no-pager directly Dec 28 12:09:26 adb shell sucks Dec 28 12:09:34 WOSQI works ;) Dec 28 12:09:56 Wifi should work from settings app Dec 28 12:10:03 Might need a few tries to connect but works Dec 28 12:10:11 the settings app won't start Dec 28 12:10:29 Hmmz it should Dec 28 12:10:32 Does for me Dec 28 12:10:41 Otherwise reboot Dec 28 12:10:52 i have. that's what started this problem. Dec 28 12:11:07 Flash latest nightly :P Dec 28 12:11:14 Solves some issues though not rebooting on TP Dec 28 12:11:24 Not sure if morphis already has some clues on that Dec 28 12:11:42 settings app doesn't start sometimes with the latest images Dec 28 12:11:48 same for just type Dec 28 12:15:03 oh.. hey.. third time it showed a window! Dec 28 12:15:24 it is somehow due to rendering problems with qtwebkit and qtwayland Dec 28 12:15:27 oooh, i love that animation between the wifi selector and the password Dec 28 12:15:56 the webview doesn't render correctly when the window is moved from invisible to visible Dec 28 12:15:57 ok no keyboard.. rebooting. lol Dec 28 12:16:39 once i get it connected to my wifi is there a way to get to it via ssh? Dec 28 12:19:14 EricBlade: no Dec 28 12:19:51 that is not integrated yet Dec 28 12:20:12 with only 1-2 people fixing those things ... it is hard :) Dec 28 12:21:09 EricBlade: when the keyboard disappears check if maliit is still up Dec 28 12:22:48 EricBlade: one of the next things will be a rndis based solution to connect with the device Dec 28 12:26:50 aha! Dec 28 12:27:02 ssh-keygen -A; sshd -p 23 Dec 28 12:27:08 er.. Dec 28 12:27:15 ssh-keygen -A; /usr/bin/sshd -p 23 Dec 28 12:27:18 for some reason 22 is occupied Dec 28 12:45:40 ok, so, i went to start mail, and everything got stupid slow, and according to top, LunaWebAppManager is hammering the memory usage Dec 28 12:47:51 yeah Dec 28 12:48:04 qtwebkit uses quite some memory Dec 28 12:48:51 not sure yet what we can do against that Dec 28 12:49:55 I would think that the memory should all be in QtWebProcess Dec 28 12:51:11 and once it's down to about 170mb free in the swapper, it crashes Dec 28 12:51:18 *it being the touchpad Dec 28 12:56:20 i'm going to guess that wam being at 61% memory consumption, webprocess at 10%, luna-next at 52% cpu and wam at 6% constantly .. is indicative of wam constantly redrawing unnecessarily .. and as to why wam is holding so much ram .. no clue. Dec 28 12:58:26 ok, killing wam causes a reboot Dec 28 13:00:02 lol @ email notifications Dec 28 13:04:04 EricBlade what image are you running? Dec 28 13:04:19 Tofe fixed some redrawing issue recently, I believe after the latest stable release Dec 28 13:04:31 stable Dec 28 13:05:04 You might want to move to nightly I think that fixed an issue in Luna-Next that was doing the redrawing all the time Dec 28 13:05:25 ok webappmanager is not releasing any appreciable amount of memory when closing an app.. and seems to be generating a huge amount of unexpected memory usage Dec 28 13:06:41 i wonder if i can build wam in windows Dec 28 13:07:00 EricBlade: Good investigation work :) Dec 28 13:07:06 Lots of things need fine tuning still Dec 28 13:07:21 Lots of stuff is made to just work, but hasn't been thoroughly tested or optimized Dec 28 13:07:31 EricBlade: the email notifications are completely white atm as we're not rendering the webview with transparency enabled which the dashboard styles assume Dec 28 13:08:11 i'm going to very quickly need to learn how to make sshd run at startup heh Dec 28 13:09:20 http://webos-ports.org/wiki/LuneOS_Enyo_App_Test_Results is what I mocked up quickly ;) Dec 28 13:13:41 we could dump some memory by not starting the inspector Dec 28 13:14:09 EricBlade: how much? Dec 28 13:14:45 if that is a good amount I will only enable it when the developer mode is enabled Dec 28 13:15:11 i don't know. probably not a huge amount. and that's all in the webprocess anyway. Dec 28 13:16:38 hm Dec 28 13:16:52 but seems like there is a real leak when windows are closed and their memory is not released Dec 28 13:17:56 normally, the inspector only really has any effect when there's clients connected to it, but i know we had at least one bug a long time ago, where having the inspector enabled caused some operations to not run through JIT Dec 28 13:18:07 or DFG or some other three letter acronym Dec 28 13:20:46 hm Dec 28 13:23:59 EricBlade: if you're right that should give us some more memory: https://github.com/webOS-ports/luna-webappmanager/pull/6 Dec 28 13:25:08 sure, that should work. if we were using upstart, i'd put it in the upstart script .. but i have no idea how systemd works. Dec 28 13:25:44 it won't hurt. as is, it looks like the inspector is being bound to localhost, so it wouldn't be possible to use it anyway Dec 28 13:26:03 at least judging from the log saying that it was starting up, and to use it, connect to 127.0.0.1:1122 Dec 28 13:27:11 the current way of using it is to forward the port through adb Dec 28 13:27:17 adb forward tcp:1122 tcp:1122 Dec 28 13:27:24 that makes it available on your local PC Dec 28 13:27:24 oooh. i see. Dec 28 13:27:37 not very intruitive but works Dec 28 13:27:56 yeah. there's a setting somewhere that can be set to have it listen on all interfaces Dec 28 13:28:15 ok Dec 28 13:28:25 not sure if we want this Dec 28 13:28:49 anyway, a brief look seems to indicate that the window close signal is deleting the WebApplicationWindow, but is not closing the QWindow Dec 28 13:30:07 that is right, but the mWindow is picked from the QQmlEngine Dec 28 13:30:13 mRootItem = mEngine.rootObjects().at(0); Dec 28 13:30:20 mWindow = static_cast(mRootItem); Dec 28 13:30:45 so I assume that the QQmlEngine cares about it's deletion once the WebApplicationWindow goes away Dec 28 13:31:31 hmm. creating a new QQmlEngine for each window? Dec 28 13:31:53 yes, that is how it is implemented atm Dec 28 13:31:54 yeah that's probalby right Dec 28 13:32:05 not sure how much overhead that is Dec 28 13:32:13 Yeah I'm not sure if it's possible to use just one Dec 28 13:32:18 but should not that big as we're not doing very much with the engine Dec 28 13:32:44 not sure if calling http://doc.qt.io/qt-5/qwindow.html#destroy improves something Dec 28 13:33:20 close() says it calls destroy Dec 28 13:34:10 I think the first thing I would do is make sure that QQmlEngine::quit is being fired when we think it is Dec 28 13:35:14 EricBlade: yeah Dec 28 13:36:37 EricBlade: btw. one other problem we have right now is: Dec 28 13:36:52 the window for the just type app is created early on startup Dec 28 13:37:09 we call QQuickWindow::show for it to register it with the compositor and the cardshell Dec 28 13:37:47 when the cardshell starts rendering the window is not visible until the user tips the just type area Dec 28 13:37:55 that brings the window into the foreground Dec 28 13:38:04 in some cases the WebView doesn't render anything Dec 28 13:38:13 mainly when another app was started before Dec 28 13:38:35 the DOM tree for the app is correctly setup (inspector shows everything) Dec 28 13:38:45 also QML parts are rendered corretly Dec 28 13:38:57 but the webview doesn't seem to detect that it can now start rendering Dec 28 13:39:31 that sounds like it could be a pretty complex interaction Dec 28 13:40:00 my current suspection is this is somehow related to the scenegraph changes done somewhere between 5.2 and 5.4 that the webview is not rendered as not visible on the screen and then can't really resume rendering once it is visible Dec 28 13:40:48 i think 5.2 removed the ability to render when not showing Dec 28 13:41:19 yeah Dec 28 13:41:29 that is fine Dec 28 13:41:38 but it should restart rendering when becoming visible Dec 28 13:42:10 yeah i'd need to understand more of how everything works than just a quick look Dec 28 13:42:56 me too Dec 28 13:43:06 just thought you might have a quick idea Dec 28 13:43:19 will ping some of the qt guys in the new year Dec 28 13:43:51 my feeling on the memory consumption problem is that unless this works vastly differently from what i am familiar with, a- lunawebappmanager is allocating a LOT of memory that should be investigated b- it's probably not destroying windows when done with them Dec 28 13:46:06 on the system i'm familiar with, webappmanager only consumes a couple of megs per app, and 16mb for each window Dec 28 13:46:23 hm Dec 28 13:46:48 is there a good way to measure the memory usage per window and app? Dec 28 13:47:47 nizovn: thanks for https://github.com/webOS-ports/location-service/pull/2 ! Dec 28 13:47:50 i'm not real sure on that. i saw it jump about 400mb when i launched the mail window Dec 28 13:47:58 finally we have positioning support back Dec 28 13:49:36 nice Dec 28 13:50:03 ok, well, i guess i'll get to work on setting up something with which i can build. Dec 28 13:50:25 yeah, would be a good start Dec 28 13:51:03 suggestions for a build system? i'll need to configure an emulator for it Dec 28 13:52:01 Ubuntu 12.04 or 14.04 Dec 28 13:52:20 is the easiest way and that is what I, others and the builders are working with Dec 28 13:53:49 hmm. i only have 500gb open. ugh. Dec 28 13:56:28 80-100gb suitable? Dec 28 13:59:15 expect 40-60g for the buil Dec 28 13:59:20 so 80-100 should be ok Dec 28 14:04:25 JaMa: our builders are still 12.04 or already 14.04? Dec 28 14:11:59 morphis: so i need to move one file location_getposition.c to location-service/tests? Dec 28 14:12:30 EricBlade: regarding memory management: we will have a memory monitor soon so we can detect situations where we don't have much memory less Dec 28 14:12:40 nizovn: hm Dec 28 14:13:01 are we intentionally rebooting when webappmanager gets killed? Dec 28 14:13:03 nizovn: I am wondering why you don't call the dbus method directly from the service and use a another process for that Dec 28 14:13:09 EricBlade: no Dec 28 14:13:36 maybe it was just coincidence then that every time i went to kill it a reboot happened right after Dec 28 14:13:39 possible that this is due to the kernel Dec 28 14:13:50 EricBlade: how do you kill it? Dec 28 14:13:56 systemctl stop LunaWebAppManager? Dec 28 14:13:58 killall -9 Dec 28 14:14:08 use systemctl instead Dec 28 14:14:16 and then a systemctl start LunaWebAppManager Dec 28 14:14:55 oh... umm Dec 28 14:15:01 ofonod is blowing me up now Dec 28 14:15:13 systemctl disable ofono Dec 28 14:15:32 that is one bug I need to fix too ... Dec 28 14:15:48 it is going crazy when it doesn't find the rild service Dec 28 14:16:15 yeah.. too many open files .. sucking up memory somewhere probably.. and disabling it doesn't seem to effectively stop it Dec 28 14:16:40 could be also possible that we're leaking kernel fds somewhere in the graphics subsystem Dec 28 14:17:20 ... and it rebooted. Dec 28 14:17:54 i killed ofonod with kill -9 because systemctl said it wasn't running Dec 28 14:18:03 about 30 seconds later, reboot Dec 28 14:20:08 ok, killing wam without memory pressure going on doesn't seem to hurt it, it just restarts Dec 28 14:26:52 dammit. the dreaded "AMD-V is not available". Dec 28 14:27:18 morphis: as i can see geoclue allow only one client for each executable. As request can take up to 20sec, it's possible to receive another request with different accuracy.(not sure how first will work if overwrite accuracy). Dec 28 15:13:38 nizovn: I see .. Dec 28 15:14:02 ok do we have current build instructions? Dec 28 15:20:03 well i found what's on the wiki, so hopefully they are :-) yay Dec 28 15:20:31 morphis: https://github.com/webOS-ports/meta-webos-ports/pull/85, https://github.com/webOS-ports/meta-webos-ports/pull/86 and https://github.com/webOS-ports/meta-webos-ports/pull/87 for the work John has done on migrating to Enyo 2.5.1 Dec 28 15:21:36 I tested on N4 and looks OK now, merged changes and did the PR for dizzy branch on meta-webos-ports above Dec 28 15:21:45 Preware and a few others are still pending Dec 28 15:26:38 Herrie: great Dec 28 15:26:48 EricBlade: they are current :) Dec 28 15:27:52 ok well i started that. going to go tkae a nap now. bbl .. then i'll try to figure out a workflow that will work for me Dec 28 15:28:50 EricBlade: They worked for me about a month ago ;) Dec 28 15:29:08 Just ordered 2 4TB disks yesterday to replace the old bunch of disks in this PC ;) Dec 28 15:29:16 Can re-organize my stuff a bit better then :P Dec 28 15:38:58 EricBlade: I added an explicit mWindow->close() in the destructor of the webapplicationwindow now, that seems to help Dec 28 15:40:46 morphis: so do i need to move file? :) Dec 28 15:45:18 nizovn: no, leave it where it is then Dec 28 15:45:38 ok Dec 28 16:11:12 nizovn: merged Dec 28 16:11:24 nizovn: can you send the PR for meta-webos-ports too? Dec 28 16:11:37 ok Dec 28 17:25:55 morphis: sorry, can Dec 28 17:26:09 't understand what is happening in testr Dec 28 17:48:00 nizovn: you ever wrote a enyo app before? Dec 28 17:49:48 nizovn: https://github.com/webOS-ports/org.webosports.app.testr/tree/master/source is where you have to look at Dec 28 17:50:04 look at this commit here: https://github.com/webOS-ports/org.webosports.app.testr/commit/a356bf5a47961cb214ce7ffcb382e3e749c5089b Dec 28 17:50:37 nizovn: but if you don't have any experience in JS/enyo I can implement or ask someone else to do that Dec 28 17:58:20 nizovn: what are the other things you will work on next? Dec 28 17:59:30 no, i wrote one mojo app and copied one stub enyo example app before. seems need to know a lot of enyo documentation Dec 28 18:01:13 :) Dec 28 18:01:14 as imaccountvalidator and imlibpurpleservice are done, i don't have pending jobs Dec 28 18:01:36 nizovn: we're missing an account validator for the im services Dec 28 18:01:46 Herrie found an enyo1 based one Dec 28 18:01:53 we already have account templates for most services Dec 28 18:02:16 see https://github.com/webOS-ports/messaging-accounts Dec 28 18:02:27 but the validator needs to be added and we need to test both services Dec 28 18:02:47 also we need to work on fixing https://github.com/webOS-ports/imlibpurpleservice/commit/8a6684e396aa9b9d2aecd179e8fc53b585331ec0 Dec 28 18:03:05 maybe that is something you're interested in Dec 28 18:05:31 we don't have the app yet to test everything but from the command line we can make sure things are in place Dec 28 18:09:23 morphis: Do we know what we want to sanitize exactly and what not? I found some examples in c++ that might be useful ;) Dec 28 18:10:21 Herrie: that is something we have to figure out Dec 28 18:10:27 but we need to remove certain tags Dec 28 18:11:22 Ah it seems that this might give a hint for IM: https://github.com/webOS-ports/imlibpurpleservice/blob/master/src/IMMessage.cpp#L49 Dec 28 18:12:07 Not sure why we wouldn't allow tags in IM messages? I think those should be optionally allowed or maybe only from known contacts? Dec 28 18:12:28 would be something we can implement Dec 28 18:12:36 but security wise it is better to block those things first Dec 28 18:17:33 morphis: so i need to find some c library to sanitize html and get to work commented lines? Dec 28 18:19:20 nizovn: Yes or write one ;) Dec 28 18:19:32 It shouldn't be too hard I guess, but then again I don't know C Dec 28 18:19:52 The IMMessage.cpp gives quite a good idea what it should do in the commented code Dec 28 18:20:50 ok Dec 28 18:31:50 nizovn: another thing you could do afterwards is mms support Dec 28 18:33:19 I have the backend ready, just the integration into the webos world is missing Dec 28 18:36:31 interesting, but what is the integration? not enyo? Dec 28 19:15:09 hello .. i don't get the topic but is there a topic for luneos for nexus4 ? Dec 28 19:16:21 jcbjoe: Ask, got a N4 myself ;) Dec 28 19:16:29 morphis & Tofe have one too ;) Dec 28 19:16:32 cool Dec 28 19:16:55 omg i haven't had a webos device in a while hahah Dec 28 19:17:03 omg i feel bad now in my history Dec 28 19:17:49 i used to have a palm pre2 Dec 28 19:18:13 i used to be in the palm pre webos-ports all that stuff before hp got a hold the whole thing went down Dec 28 19:18:19 i fell off lol Dec 28 19:18:28 I see Dec 28 19:18:50 luneos looks nice .. can it be used as a daily driver ? Dec 28 19:19:00 Well we have a LuneOS (Open webOS Fork + improvements) for N4 + HP Touchpad. It's stil alpha and not daily driver ready yet;) Dec 28 19:19:13 But it's good enough to give you an idea of what will be ;) Dec 28 19:19:13 oh ok Dec 28 19:19:40 We have no phone + messaging working yet (well on command line but it's not fully integrated yet), that's currently in the works Dec 28 19:20:53 HP released limited bits as open source, some were not even usefull and were dropped in favor of proper open source replacements. Other parts had to be completely rewritten from scratch Dec 28 19:21:19 So as you can imagine that's a big task for a small team of devs ;) We don't have the 100's Palm & HP used to have ;) Dec 28 19:22:45 and of course no ETA Dec 28 19:22:49 1 rule i konw right Dec 28 19:24:40 Yeah;) Dec 28 19:25:34 Also we currently require Android 4.2.2 on the N4 due to some firmware for the hardware bits in the phone (will be updated probably shortly as well, but we had some issues with 4.4.x we still need to tackle) Dec 28 19:26:04 It might be possible to multi-boot using MultiRom (http://forum.xda-developers.com/nexus-4/orig-development/mod-multirom-v24-t2472295) but I haven't got around to playing with that myself yet Dec 28 19:26:26 whats your current phone ? Dec 28 19:27:43 I have a HP Veer as daily driver, it's getting a bit slow, but does what I need in general ;) Dec 28 19:27:54 wow a veer Dec 28 19:28:00 in 2014 ? Dec 28 19:28:02 Virtual Keyboards are a pain for me :P Dec 28 19:28:04 :P Dec 28 19:28:32 Everything still works more or less.. Some things are broken, but I mainly call, email, sms, whatsapp, twitter and that all works still Dec 28 19:28:56 yep true multitasking Dec 28 19:43:38 Herrie: reading up a bit aren't they on alpha 2 ? Dec 28 19:44:04 wow that was back in 26 June 2013 Dec 28 19:48:34 Yeah ;) Dec 28 19:48:47 We have monthly rolling releases now Dec 28 19:49:08 After alpha 2 we decided to rewrite UI from scratch because old one was a no go basically Dec 28 19:49:31 And renamed it to LuneOS ;) Dec 28 19:50:10 We have had 4 monthly releases now, 5th one due soon in early January Dec 28 20:43:47 morphis: You know if there's a way to have the gh-pages branches of the Enyo apps to automatically update to latest stuff from master branch when we do a commit there? Dec 28 20:44:54 I just updated all Enyo apps GH-pages branch to latest commits from master, but would be good if it's something automated instead ;) Dec 28 20:45:04 Or we should just add it in our pre/post release todo list ;) Dec 28 21:04:27 Herrie: no, don't know Dec 28 21:22:16 ... oh, awesome. apparently Virtualbox doesn't look like a CPU load to Windows, so it suspended a few minutes after i went away. Dec 28 21:24:33 or maybe windows only goes on interactive activity. Dec 28 21:24:56 but i haven't had it suspend on me in several days. :-S so no clue. oh well. going now. Dec 28 21:25:16 EricBlade: I played a bit with releasing window resources from our webappmanager today Dec 28 21:25:31 on the N4 things seem to go well Dec 28 21:25:58 it decreases memory always when windows are closed Dec 28 21:26:29 so some platform specific blah blah .. ? Dec 28 21:26:35 think so Dec 28 21:27:47 need to check this on the tp Dec 28 21:28:28 so.. what is platform specific? Dec 28 21:29:40 the whole android graphics shit Dec 28 21:30:17 look, the tp kernel was ported from qcom open source parts to work with binaries coming from somewhere which suite what the TP needs Dec 28 21:30:30 now we're coming and using it again differently Dec 28 21:31:02 if you look at the dmesg output you will very quickly see also warnings saying that ION has no more space Dec 28 21:31:40 what is ion ? Dec 28 21:31:50 that all really smells like that there is something eating more memory than it should and doesn't free it and that down the stack where the graphics management happens Dec 28 21:32:00 ION is the kernel memory allocator for graphics in Android Dec 28 21:32:19 http://lwn.net/Articles/480055/ Dec 28 21:32:49 basically a generic approach to have the same interface for different vendors Dec 28 21:37:21 that looks a bit outside of my domain, as far as debugging it might go. Dec 28 21:40:17 EricBlade: yeah, even for me but someone needs to go down and try to find out what is wrong Dec 28 21:43:27 morphis: 12.04 Dec 28 21:43:48 JaMa: didn't we upgrade to 14.04 some time back? Dec 28 21:44:09 only jenkins master Dec 28 21:44:29 morphis, JaMa do we need to upgrade the builder(s)? Dec 28 21:45:48 I don't know about any pressing reason Dec 28 21:45:55 \o/ Dec 28 21:46:47 ka6sox: yeah, no pressing reason Dec 28 21:46:51 JaMa: ah ok Dec 28 21:47:02 just remembered ka6sox was updating one of the servers Dec 28 21:47:09 just the master Dec 28 21:47:25 slave builders are still 12.04 Dec 28 21:49:22 ka6sox: ok Dec 28 21:50:06 ok how do i stop a service from running? Dec 28 21:51:49 service foo stop Dec 28 21:52:11 service: command not found Dec 28 21:55:57 no Dec 28 21:56:03 systemctl stop Dec 28 21:56:07 we're on systemd Dec 28 21:56:24 EricBlade: http://www.freedesktop.org/software/systemd/man/systemctl.html Dec 28 21:56:39 ty Dec 28 22:00:59 hmm. running valgrind on lunawebappmanager doesn't work so well, ls2 blocks the connection... anyone remember how to turn off ls2 security? Dec 28 22:02:10 yes Dec 28 22:02:24 look at /etc/ls2/ls-public... Dec 28 22:02:32 there is a Security=true switch Dec 28 22:02:50 change it to false and reboot Dec 28 22:03:46 ah don't forget to do the same for the private bus too Dec 28 22:13:08 EricBlade: so you're running valgrind on the webappmanager now? Dec 28 22:13:39 yeah Dec 28 22:13:53 probably won't tell me a whole lot without debug symbols though Dec 28 22:14:06 yeah, you have to install them Dec 28 22:14:23 JaMa: if I remember well there was a switch to build an image with all symbols included, right? Dec 28 22:16:19 umm.. is there a seperate instance of lunawebappmanager per app? o.o Dec 28 22:16:24 wassup Dec 28 22:17:24 i started and closed notes, and valgrind terminated.. the numbers look grim. Dec 28 22:18:20 definitely lost 1k indirectly lost 50k possibly lost 226k .. 591988 errors. Dec 28 22:18:34 with 897986 errors suppressed. Dec 28 22:18:48 psh, that's barely any errors Dec 28 22:18:52