**** BEGIN LOGGING AT Mon Dec 29 02:59:58 2014
Dec 29 04:04:41 <EricBlade> package lxc fails to build
Dec 29 04:04:55 <EricBlade> | error: Doxyfile not found and no input file specified!
Dec 29 04:32:05 <EricBlade> adding a dummy Doxyfile to the dir /home/eric/builds/lune/webos-ports-env/webos-ports/tmp-glibc/work/armv7a-vfp-neon-webos-linux-gnueabi/lxc/1.0.1-r0/build/doc/api .. got it to at least install
Dec 29 04:32:12 <EricBlade> i'm sure that's not the right way :D
Dec 29 07:14:35 <EricBlade> hooray! build complete. now how to get it to the touchpad ?
Dec 29 07:14:50 <EricBlade> and preferably including debugs for qt and lunawebappmanager
Dec 29 07:32:42 <Herrie> EricBlade: Not sure where your Doxyfile error comes from. I only tried N4 build tbh
Dec 29 07:33:03 <Herrie> It usually fails first time, then I re-run it and it continues and completes successfully
Dec 29 08:08:27 <EricBlade> Herrie|2: i don't know. there's clearly no Doxyfile in that dir, and it clearly wants one.
Dec 29 08:12:26 <EricBlade> Herrie|2: now.. how to get the debug files i want and my image over to my touchpad ... ?  do we have any short methods of doing it, or just adb copying everything
Dec 29 08:12:36 <EricBlade> i see that sftp doesn't work
Dec 29 08:13:29 <EricBlade> scp doesn't deal well with multiple files and paths and stuff easily
Dec 29 08:13:37 <EricBlade> neither does adb
Dec 29 08:15:55 <JaMa> morphis: yes, dbg-pkgs to IMAGE_FEATURES
Dec 29 08:30:48 <EricBlade> doesn't that kind of create a multi gigabyte image?
Dec 29 08:30:56 <EricBlade> like many many gigabytes
Dec 29 08:52:56 <JaMa> it's a lot bigger, but IIRC should fit in
Dec 29 09:00:24 <EricBlade> i guess i'll give it a shot and see what cooks up
Dec 29 09:09:30 <JaMa> shouldn't take long, it just creates new rootfs with more packages in it (no rebuilds)
Dec 29 10:41:52 <Herrie|2> nizovn: I came across this one: Might be of help for the HTML sanitizing? https://github.com/google/gumbo-parser
Dec 29 10:42:48 <nizovn> thanks, do you know is message guarantied to be valid?
Dec 29 10:46:28 <morphis> nizovn: you mean the incoming one?
Dec 29 10:46:44 <nizovn> yes
Dec 29 10:47:08 <morphis> the incoming one can contain everything
Dec 29 10:47:27 <morphis> that is why we have to remove things we don't trust
Dec 29 10:48:29 <nizovn> then seems it's necessary to use something like tidy
Dec 29 10:49:47 <morphis> yeah
Dec 29 15:27:45 <EricBlade> JaMa: looks like 1.6gb compressed
Dec 29 15:32:25 <EricBlade> gzip -l on the file says uncompressed it's 100mb, which clearly is not correct, so i chcked the long way.. and that tells me 4.4gb
Dec 29 15:34:01 <EricBlade> holy wtf. i just booted legacy webos on it, in under 15 seconds or so.
Dec 29 15:40:22 <EricBlade> hmm. tailor can't resize an android partition? ugh
Dec 29 15:48:12 <EricBlade> can i resize the android volume from command line?
Dec 29 15:49:23 <Herrie|2> You should be able to
Dec 29 15:49:30 <Herrie|2> Let me dig up that link
Dec 29 15:51:15 <Herrie|2> http://www.webos-internals.org/wiki/UbuntuChroot#Alternative_Step_1:_Re-partitioning_by_hand
Dec 29 15:51:39 <Herrie|2> Should give you some pointers to do it manually...
Dec 29 16:07:54 <EricBlade> Herrie|2: hmm. i did a lvresize on it to 6.8gb, and it said it was successful but mounting it only shows 2gb still
Dec 29 16:10:47 <EricBlade> maybe a mke2fs on it
Dec 29 16:11:03 <nizovn> or resize2fs
Dec 29 16:11:19 <EricBlade> probably that too
Dec 29 16:11:42 <EricBlade> yeah mke2fs fixed the size. yay.
Dec 29 16:15:43 <EricBlade> ok i really need to figure out a better workflow than "build thing on laptop, move it to desktop, boot webos, mount rootfs, go to usb mode, copy thing to touchpad, go to command line, untar it, reboot" .. heh
Dec 29 16:16:18 <Herrie|2> EricBlade: Yeah that's not a very nice workflow :P
Dec 29 16:16:35 <Herrie|2> That's how I upgrade currently too, not very efficient ;)
Dec 29 16:18:17 <EricBlade> hmm. doesn't look like anyone has ever bothered to implement wifi nfs booting
Dec 29 16:23:16 <EricBlade> at least not without some pretty drastic measures. oh well
Dec 29 16:25:45 <EricBlade> well getting sshd to run on boot will be my first victory i think
Dec 29 16:26:12 <EricBlade> looks like sftp isn't included in the default, not sure how difficult that would be to add
Dec 29 16:26:36 <EricBlade> i do see that i can mount LuneOS via USB, but I'm not sure what I'm accessing there
Dec 29 16:28:08 <Herrie|2> I think just /media/internal
Dec 29 16:28:18 <Herrie|2> Which is shared with webOS AFAIK
Dec 29 16:28:55 <EricBlade> what i got was a 2gb partition that had 1.6gb free, so it was apparently the LuneOS root, but there were no files visible in it
Dec 29 16:38:31 <EricBlade> hmm. i wasn't expecting large quantities of source code in the debug image. is that normal?
Dec 29 16:42:29 <Herrie|2> Dunno tbh
Dec 29 16:42:52 <Herrie|2> Where are you seeing source code? For QML apps we have the uncompiled QML code visible ;)
Dec 29 16:47:33 <EricBlade> a ton of header files and some source code are in /usr/lib/.debug/*
Dec 29 16:55:09 <Herrie|2> Hmmz could be, I don't touch C(++) stuff :P
Dec 29 16:59:09 <EricBlade> ok that took somewhere north of forever to uncompress the image
Dec 29 17:02:00 <EricBlade> oh.. hmm. not good. reboot loop. do i need to replace the kernel?
Dec 29 17:03:03 <EricBlade> oh.. it only looped twice
Dec 29 17:04:43 <EricBlade> i wonder what the cause is of the very slight delay between touch input and response.
Dec 29 17:07:11 <Herrie|2> Feel free to debug
Dec 29 17:07:21 <Herrie|2> I'm sure there's ton of optimizations that need to take place
Dec 29 17:07:37 <Herrie|2> It already runs a lot smoother compared to earlier, but it's still not hw accelerated AFAIK
Dec 29 17:07:41 <Herrie|2> So all is done by software
Dec 29 17:07:55 <Herrie|2> And the TP CPU speed is bottleneck a bit there I guess. N4 is smoother
Dec 29 17:13:13 <EricBlade> i wonder if it's not passing in the tap until after the splash effect is played
Dec 29 17:14:52 <Herrie|2> Could also be, just from my experience TP is a bit more laggy compared to N4
Dec 29 17:14:58 <Herrie|2> But that's no wonder seeing the diff in hardware
Dec 29 17:15:49 <EricBlade> other than the performance of the emmc being very poor compared to modern hardware, the touchpad really isn't that bad
Dec 29 17:16:41 <EricBlade> sure, sure, top of the line things have faster cpus and more cores, and more ram.. but 1.5ghz dual ought to be more than enough :-)
Dec 29 17:17:30 <Herrie|2> TP is quite OK seeing it's 3'5 years old
Dec 29 17:19:00 <EricBlade> this looks really not good, but it doesn't seem to be fatal: http://pastebin.com/c7K4CzVa
Dec 29 17:20:04 <EricBlade> and it really sucks that this is a hardware specific issue on wam .. if this occurred in emulator, it would be much faster to debug potentially
Dec 29 17:22:39 <Herrie|2> Tofe: ping
Dec 29 17:23:19 <EricBlade> splash screen comes up all bugged out when running in valgrind
Dec 29 17:27:16 <EricBlade> ok if anyone else wants to look at this... https://www.dropbox.com/s/sw0upg33snkdicz/valgrind-notes.txt?dl=0
Dec 29 17:27:22 <nizovn> EricBlade: reticle was my first commit here:) seems ok to me: https://github.com/webOS-ports/luna-next/commit/5756d6680df511f134c50b43c146ea20ec69c202
Dec 29 17:32:05 <nizovn> btw on emulator there is another issue: LunaWebAppManager cause 100% cpu load somewhere on futexes.
Dec 29 17:41:26 <EricBlade> hmm. i think hte problem i'm debugging might've gone away
Dec 29 17:42:25 <EricBlade> or maybe it's gone away because LunaWebAppManager is closing after every app close.
Dec 29 17:45:57 <EricBlade> and here's a run on email: https://www.dropbox.com/s/hlr0swt39v6cvkf/valgrind-email.txt?dl=0
Dec 29 17:48:34 <EricBlade> scrolling doesn't work right in Email, but other than that, the memory problem seems to be gone in current dev
Dec 29 17:49:00 <EricBlade> only using 21% on lunawebappmanager with notes and mail running
Dec 29 17:51:14 <EricBlade> ok, nevermind, closing windows didn't crash it now.
Dec 29 17:51:22 <EricBlade> and it's still using 34% after three windows open
Dec 29 17:52:50 <EricBlade> ok so it's still not freeing memory after closing a window.. but it's not running absolutely rampant with memory allocation like the stable was
Dec 29 17:53:47 <EricBlade> and it's really quite fast until it starts getting deep into swapper territory.  but since it's not constantly hitting the swapfile anymore it stabilizes some once it does
Dec 29 17:56:48 <EricBlade> takes sooooo long to start lunawebappmanager in valgrind with full debug symbols
Dec 29 18:02:08 <EricBlade> nizovn: the reticle does seem to get a little off sometimes, perhaps rounding the coords off isn't good? not sure.
Dec 29 18:05:39 <EricBlade> not sure i can come up with anything even remotely concrete with it crashing out as soon as i close a window when running in valgrind
Dec 29 18:18:56 <EricBlade> oh, hey. that was easy. systemd is already configured to deal with ssh
Dec 29 18:18:56 <EricBlade> systemctl enable sshd.socket
Dec 29 18:20:08 <EricBlade> first time you connect it takes a very long time, but after that it's quick.
Dec 29 18:21:56 <nizovn_> hm.. on touchpad iirc there is some problem with touchscreen - at least if touch it very quickly, then phantom clicks will appear on screen edges.
Dec 29 18:22:25 <EricBlade> nizovn not that drastic, just seems like it's a little bit off sometimes in the upper half of the screen
Dec 29 18:23:11 <EricBlade> i wonder if a tiny bit of movement triggers a change in the point (so it's not first() ) without triggering an Update
Dec 29 18:23:51 <EricBlade> or maybe it's just my eyeballs
Dec 29 18:26:24 <nizovn_> ok will check
Dec 29 18:35:43 * DougReeder waves hello
Dec 29 18:48:35 <EricBlade> gm DougReeder
Dec 29 18:49:32 <DougReeder> Anything important happen while I was AFK this past week?
Dec 29 18:52:08 <EricBlade> i got Lune installed and tried to valgrind my way to figure out what's going wrong with luna-wam on touchpad.
Dec 29 18:52:58 <EricBlade> looks like significant gains in memory and cpu have been achieved since the last stable release
Dec 29 18:54:00 <EricBlade> but i'm not any closer to figuring out what's going wrong with luna-wam on touchpad. i think some other things may need to get investigated first, like why does luna-wam crash out after closing an app when running in valgrind, or apparently in other situations as well
Dec 29 18:55:18 * DougReeder nods
Dec 29 19:19:45 <morphis> EricBlade: I never looked how the memory layout on the touchpad is orgnized
Dec 29 19:20:32 <morphis> so it's possible that we run out of memory cause the RAM is full or the GPU eats too much memory
Dec 29 19:47:03 <morphis> DougReeder: ping
Dec 29 19:47:12 <DougReeder> pong
Dec 29 19:47:35 <morphis> DougReeder: you have time to add another test case to the testr app?
Dec 29 19:47:41 <DougReeder> Sure.
Dec 29 19:47:57 <morphis> you will not be able to test the ls2 calls yet
Dec 29 19:48:05 <morphis> just one
Dec 29 19:48:14 <morphis> we have a brand new location service
Dec 29 19:48:17 <morphis> https://github.com/webOS-ports/location-service
Dec 29 19:48:22 <DougReeder> Nifty!
Dec 29 19:48:25 <morphis> I want a simple test case in testr to determine the current position
Dec 29 19:48:35 <morphis> the current implementation in the latest nightly should give you a static position
Dec 29 19:49:11 <DougReeder> Does it have the same API as the webOS location service?
Dec 29 19:49:25 <morphis> it follows the API at https://developer.palm.com/content/api/reference/services/gps.html
Dec 29 19:49:30 <morphis> yes
Dec 29 19:49:33 <morphis> nizovn implemented it
Dec 29 19:49:37 <morphis> ask him if you have any more questions
Dec 29 19:49:52 <DougReeder> Great! So, calls should suceed under webOS?
Dec 29 19:49:56 <morphis> afaik accuracy is the only argument supported for getCurrentPosition atm
Dec 29 19:49:57 <morphis> yeah
Dec 29 19:50:01 <DougReeder> As, in, get the real position?
Dec 29 19:50:24 <morphis> in LuneOS we're not that far yet
Dec 29 19:50:38 <morphis> it will only determine your position based on WiFi networks around you
Dec 29 19:51:26 <DougReeder> You want a test which gets a single location.  A test succes is just that a valid location is returned?
Dec 29 19:53:14 <morphis> yeah, just getting the position and display the result
Dec 29 19:53:21 <morphis> that all when the user presses a button
Dec 29 19:54:25 * DougReeder nods
Dec 29 19:55:27 <Herrie|2> Morphis is it correct there are no test cases for telephony yet?
Dec 29 19:55:46 <Herrie|2> The menu option is there but nothing when you click on it
Dec 29 19:55:47 <morphis> Herrie|2: yeah
Dec 29 19:56:12 <Herrie|2> Seems we have a problem with Enyo 2.5.1 Input fields and VKB
Dec 29 19:56:19 <Herrie|2> It acts weird in Settings and Memos app
Dec 29 19:56:27 <Herrie|2> You can enter 1 char and it hides
Dec 29 19:56:42 <DougReeder> Joy.
Dec 29 19:56:54 <Herrie|2> Not sure what's the cause of it
Dec 29 19:57:17 <Herrie|2> But noticed it with both the Wifi PW input field and memo title+body
Dec 29 19:58:38 <Herrie|2> I suspect it might have to do with the change from resized into render like in https://github.com/webOS-ports/org.webosports.app.settings/commit/fb4d9934139aaff9753b4e91de8ae3c2f653777d
Dec 29 19:58:53 <Herrie|2> But not sure, also asked John aka CobbleMonster to look into it....
Dec 29 20:04:14 <morphis> Herrie|2: sounds like a regression in enyo itself
Dec 29 20:08:35 <Herrie|2> morphis: Could be not sure
Dec 29 20:10:21 <Herrie|2> I can bump to 2.5.3 pre 5 locally to see if that sorts it
Dec 29 20:15:40 <Herrie|2> Same with 2.5.3 Pre 5 :(
Dec 29 20:31:01 <EricBlade> i haven't done any looking, but what's the deal with Container.qml i think it is? i saw a bunch of runtime errors in it
Dec 29 20:37:05 <morphis> EricBlade: the ApplicationContainer wraps the webview
Dec 29 20:37:16 <morphis> it is used for nowindow and normal windowed cards
Dec 29 20:37:35 <morphis> when a card is windowed the ApplicationContainer is wrapped into a QQuickWindow
Dec 29 20:37:55 <morphis> when it is a nowindow card it doesn't use a QQuickWindow as it will be never visible
Dec 29 20:55:33 <EricBlade> huh. not getting any errors from it now.
Dec 29 21:54:19 <Herrie|2> morphis: OK did some further testing 2.4.0 is OK, 2.5.0 and up is broken
**** ENDING LOGGING AT Tue Dec 30 02:59:58 2014