2016-07-21 00:04:04	SylvieLorxu	Never ran into that one, did you run ". ./prepare-build <devicename>" first? (with <devicename> being your device, I think you used my watch so that's dory)
2016-07-21 00:04:32	itsme_	yes, i did
2016-07-21 00:04:42	itsme_	which os do you use?
2016-07-21 00:05:25	SylvieLorxu	I currently run Gentoo and Debian and I didn't have it happen on either of both
2016-07-21 07:23:41	@kido	itsme__: are you in the build directory ? Did you *source* the file with a dot before ./ instead of executing it ?
2016-07-21 11:01:00	SylvieLorxu	Hey, which language does everyone speak here? I personally speak English, Dutch and reasonable German (although German is not my mother tongue and I can't write it perfectly)
2016-07-21 11:01:50	SylvieLorxu	I suppose I'll just write a Dutch translation because we need one for testing
2016-07-21 11:01:52	SylvieLorxu	But yeah
2016-07-21 11:01:59	itsme____	I speak English, German and can write Latin?
2016-07-21 11:02:39	SylvieLorxu	Okay, then I'll just not touch German translations. Will still have to set this whole system up anyway but yeah
2016-07-21 11:03:57	@kido	I'll be in charge of the french translation
2016-07-21 11:04:40	SylvieLorxu	Okay, I'll probably start off making the Settings app both English and Dutch
2016-07-21 11:04:53	SylvieLorxu	So we have a bit of a start and can see how translations should be done for other apps too
2016-07-21 11:05:07	@kido	perfect
2016-07-21 11:05:41	SylvieLorxu	I'll probably hack on that most of today, I'll see when I'll manage to get it working :P
2016-07-21 11:07:06	@kido	good, ping me if you have any question
2016-07-21 11:07:20	SylvieLorxu	I will :)
2016-07-21 11:10:07	itsme_____	When i get it running i'll make the German translations
2016-07-21 11:11:18	SylvieLorxu	I'm personally still hoping to make translations be managed through something like Weblate or whatever system kido found that I forgot the name of. Weblate offers a free hosted instance for Free Software projects, although we could also host our own
2016-07-21 11:11:45	SylvieLorxu	I'd rather not force translators to use Qt Linguist because, frankly, it's not a nice tool to work with
2016-07-21 11:12:01	@kido	yep, I haven't decided the exact infrastructure yet but I would like to have a web tool too
2016-07-21 11:14:07	SylvieLorxu	I have personally used Transifex, CrowdIn and Weblate, of which Weblate was nicest to use, both because it's libre and it's really designed to stay out of your way. No experience with other tools though, so I can't compare between Weblate and what you found
2016-07-21 11:19:34	SylvieLorxu	If we use Weblate we should know that if we use their instance we can only use it to translate up to 25 "components" (aka: repos) https://weblate.org/en/hosting/
2016-07-21 11:25:57	@kido	that should be enough
2016-07-21 12:10:58	SylvieLorxu	Hmm, the device has been on the splash screen for several minutes now, although it does turn the screen off and I need to tap it to turn it back on, it doesn't seem to be progressing
2016-07-21 12:11:08	SylvieLorxu	How do I debug this? :(
2016-07-21 12:15:39	SylvieLorxu	Oh I have an adb shell
2016-07-21 12:17:07	@kido	yep, you have screen locking because mce is running
2016-07-21 12:17:18	@kido	if the splash screen is still on asteroid-launcher isn't running
2016-07-21 12:17:24	@kido	it is ran as part of the user@1000 service
2016-07-21 12:18:58	SylvieLorxu	Hmm...
2016-07-21 12:19:05	SylvieLorxu	Jul 21 12:09:58 dory systemd[442]: asteroid-launcher.service: Main process exited, code=killed, status=6/ABRT
2016-07-21 12:19:09	SylvieLorxu	Is the first issue
2016-07-21 12:19:16	SylvieLorxu	I suppose my change to the service file broke it
2016-07-21 12:19:50	SylvieLorxu	I only added this line though: https://github.com/TheLastProject/meta-asteroid/blob/master/recipes-asteroid/asteroid-launcher/asteroid-launcher/asteroid-launcher.service#L9
2016-07-21 12:20:22	SylvieLorxu	And that locale.conf file does exist, although it clearly wasn't read by systemd as LANG is empty
2016-07-21 12:21:05	@kido	mmhh I really don't know
2016-07-21 12:23:27	SylvieLorxu	Oh what the hell
2016-07-21 12:23:41	SylvieLorxu	I just rebooted and now it works fine
2016-07-21 12:26:50	SylvieLorxu	Hmm, I think this is a change you made: "Jul 21 12:25:40 dory lipstick[507]: qrc:/main.qml:19:1: module "QtQuick.Layouts" is not installed" when trying to open asteroid-settings
2016-07-21 12:27:09	SylvieLorxu	I thought I remember you wanted to get rid of QtQuick.Layouts, at least
2016-07-21 12:27:17	SylvieLorxu	But main.qml in asteroid-settings still uses it
2016-07-21 12:35:49	@kido	but you should have QtQuickControls installed because of asteroid-calendar
2016-07-21 12:36:02	@kido	don't you ?
2016-07-21 12:36:11	@kido	you cab build it and push it manually
2016-07-21 12:36:14	SylvieLorxu	asteroid-calendar is broken too, complains about lacking QtQuickControls
2016-07-21 12:36:49	SylvieLorxu	Which is weird seeing the recipe, hmm
2016-07-21 12:41:44	itsme___	So, build is working under Ubuntu 14.04
2016-07-21 12:46:38	SylvieLorxu	On a related note, either systemd is not reading my EnvironmentFile, despite its existance, or running "printenv" just doesn't show the value to me (neither under root nor under ceres)
2016-07-21 12:50:13	@kido	sure, printenv from a shell won't show you the environment on which the graphical session is ran
2016-07-21 12:50:33	@kido	you need to check a getenv in a user app
2016-07-21 12:50:53	SylvieLorxu	Oh pfew
2016-07-21 12:50:58	@kido	but actually this will probably not work under apps neither
2016-07-21 12:51:15	@kido	I can check it myself later if you want
2016-07-21 12:51:32	@kido	itsme___: cool :)
2016-07-21 12:52:14	SylvieLorxu	Well, sure, I'll first need to figure out why the settings (and calendar) app won't run anymore. Hopefully itsme___ can reproduce that in a bit because I can't imagine any of my changing causing it :P
2016-07-21 12:54:12	itsme___	I'll take a look on it. But I think the build in my vm will take time...
2016-07-21 12:54:44	SylvieLorxu	It takes about 4 hours or so for me too
2016-07-21 13:09:08	@kido	SylvieLorxu: yep it will be reproduced, as I told you qtquickcontrols doesn't get installed
2016-07-21 13:09:35	SylvieLorxu	kido: Oh, I thought you said it should have been installed due to asteroid-calendar
2016-07-21 13:09:52	@kido	https://github.com/AsteroidOS/meta-asteroid/blob/master/recipes-asteroid/asteroid-calendar/asteroid-calendar_git.bb
2016-07-21 13:10:05	@kido	I explained you yesterday that DEPENDS is build dependencies
2016-07-21 13:10:17	@kido	And RDEPENDS is runtime deps
2016-07-21 13:11:06	SylvieLorxu	Oh, like that
2016-07-21 13:11:32	@kido	so I had to add a RDEPENDS_${PN} += "qtquickcontrols-qmlplugins"
2016-07-21 13:11:45	@kido	once again that's my fault hehe :)
2016-07-21 13:12:00	@kido	and itsme___ will need that too
2016-07-21 13:12:37	SylvieLorxu	For both asteroid-calendar and asteroid-settings, right?
2016-07-21 13:28:51	SylvieLorxu	Okay, just adding that and recompiling a single app seems to not fix that :(
2016-07-21 13:35:59	SylvieLorxu	Bitbake is... pretty confusing if you're new to it
2016-07-21 13:38:11	@kido	SylvieLorxu: of course, this RDEPENDS means that installing the asteroid-calendar and asteroid-settings packages needs the qtquickcontrols-qmlplugins packages
2016-07-21 13:38:20	@kido	if you push another app's image directory you won't get the dependencies
2016-07-21 13:38:25	SylvieLorxu	Aww :(
2016-07-21 13:38:40	@kido	so you either need to re-push a new rootfs or just push the image/ directory of qtquickcontrols
2016-07-21 13:38:53	@kido	and that will be fine
2016-07-21 13:39:15	@kido	but the next time you build an image, qtquickcontrols-qmlplugins will be installed so the image will work fine
2016-07-21 13:39:46	SylvieLorxu	I see
2016-07-21 13:43:33	SylvieLorxu	Sweet, it doesn't crash anymore :D
2016-07-21 13:46:08	@kido	:)
2016-07-21 15:34:01	SylvieLorxu	kido: I seem to be making serious progress, I can see systemd loads the locale setting on boot because the clock changes ("Thu." becomes "do.." in German, which means we'll need to make slight changes to the watchfaces but I'll get to that later). However, Qt keeps seeing the current locale as C. Seeing how `locale` is not a valid command on AsteroidOS I am kinda lost in my attempt to figure out if there
2016-07-21 15:34:02	SylvieLorxu	are maybe no locale packages installed and searching for things like locale on http://layers.openembedded.org/layerindex/branch/master/layers/ also returns nothing. Any pointers? ^^"
2016-07-21 15:35:13	@kido	what do you mean by "Qt keeps seeing the current locale as C" ?
2016-07-21 15:35:47	@kido	asteroid-launcher's watchface's internationalization is detected by Qt, with the envvar of asteroid-launcher, which loads the config file which sets LOCALE
2016-07-21 15:36:14	@kido	however if you are talking about the QML apps, they are run in a bit special way which in the end makes them unaware of asteroid-launcher's environment
2016-07-21 15:36:51	@kido	indeed, the apps use "invoker" from mapplauncher https://github.com/AsteroidOS/asteroid-settings/blob/master/asteroid-settings.desktop#L5 to cache QML components and speed up the app starting time
2016-07-21 15:37:20	@kido	which means that apps are just libs dynamically loaded when needed by a pre-loaded process already running
2016-07-21 15:38:14	SylvieLorxu	Ah, then that is probably the issue
2016-07-21 15:38:28	@kido	https://github.com/AsteroidOS/meta-asteroid/blob/master/recipes-nemomobile/mapplauncherd/mapplauncherd-booster-qtcomponents/booster-qtcomponents-qt5.service#L9
2016-07-21 15:38:30	SylvieLorxu	Because the clock gets translated, but in asteroid-settings Qt.locale() thinks it's C, not the correct locale
2016-07-21 15:38:59	@kido	yep, because the apps run with the environment of booster-qtcomponents-qt5 which isl oaded by this service and doesn't inherit the envvars from asteroid-launcher
2016-07-21 15:39:06	SylvieLorxu	Okay, let's see if adding the language EnvironmentFile to that will work
2016-07-21 15:39:15	@kido	I'm sure it will :)
2016-07-21 15:39:16	 *	SylvieLorxu nods
2016-07-21 15:39:22	@kido	good job!
2016-07-21 15:49:28	SylvieLorxu	Unfortunately that's not it, Qt.locale().name still returns "C" in asteroid-settings. Hmm
2016-07-21 15:51:20	@kido	if you restarted the session that should be fine though
2016-07-21 15:52:02	SylvieLorxu	I did, I even tried rebooting completely, but nope :(
2016-07-21 15:52:53	@kido	well maybe it's even another process
2016-07-21 15:53:12	@kido	that would be cool if this could be applied to the whole systemd session and not only a list of services
2016-07-21 15:53:24	@kido	I wonder if we can load an EnvironmentFile for user@1000.service
2016-07-21 15:54:29	@kido	https://wiki.archlinux.org/index.php/Systemd/User#Environment_variables
2016-07-21 16:10:13	SylvieLorxu	kido: Using #3 on that list makes it propagate to apps correctly, Qt.locale().name now correctly shows "en_GB" or "nl_NL". Great! The translation is somehow still not showing up, but we're getting closer every time :D
2016-07-21 16:13:49	SylvieLorxu	Oh, seems I need to enable translation in the app's CPP code lol
2016-07-21 16:19:37	SylvieLorxu	"error: aggregate 'QTranslator translator' has incomplete type and cannot be defined"
2016-07-21 16:19:52	SylvieLorxu	I have... no C++ knowledge at all, I'm just going to assume this is a dependency thing
2016-07-21 16:21:05	SylvieLorxu	nvm, missed an include
2016-07-21 16:22:24	@kido	great! :D
2016-07-21 16:22:52	SylvieLorxu	Still doesn't work though lol
2016-07-21 16:23:04	SylvieLorxu	But I am pretty sure I'm really close now
2016-07-21 16:29:35	@kido	this is probably not a matter of changing the C++ code since you hadn't to modify asteroid-launcher
2016-07-21 16:31:12	SylvieLorxu	Well, #6 on https://byteplumbing.net/2014/01/translating-ui-texts-in-jolla-applications/ just seemed really promising, but I can't get that to work either
2016-07-21 16:31:40	SylvieLorxu	Honestly, I'm just trying random things now
2016-07-21 16:43:30	SylvieLorxu	Hmm, I should probably push the translation files to /usr/share/translation
2016-07-21 16:43:37	SylvieLorxu	*/usr/share/translations/
2016-07-21 16:43:45	SylvieLorxu	I forgot to send them to the device in the first place
2016-07-21 16:43:47	SylvieLorxu	Hahahaha
2016-07-21 16:46:29	@kido	:)
2016-07-21 16:46:43	@kido	does it work now?
2016-07-21 16:47:47	SylvieLorxu	Yup: https://lut.im/kvVv7yos4d/rSEWSZ1KZCYP8ywo.jpg
2016-07-21 16:48:21	@kido	fantastic, I'll be waiting for your pull requests :)
2016-07-21 16:50:29	SylvieLorxu	I'm going to clean up a bit and see if I can fix the build script to automatically install the languages to the correct place too, as soon as that's done I'll send a pull request to both meta-asteroid and asteroid-settings :)
2016-07-21 16:51:29	@kido	nice
2016-07-21 16:51:51	@kido	we need to find a solution for the asteroid-settings UI though
2016-07-21 16:52:55	SylvieLorxu	Probably just needs to be wrapped in some element that allows scrolling
2016-07-21 16:54:02	@kido	well, I wouldn't like a scrollable grid
2016-07-21 16:54:12	@kido	we could switch to the round screens UI
2016-07-21 16:54:39	@kido	which is already a list
2016-07-21 16:55:10	 *	SylvieLorxu compiles the round screen version quickly to see
2016-07-21 16:55:30	@kido	it's a been a while I haven't tested it, I might have broken the text margins with a recent commit
2016-07-21 16:55:38	SylvieLorxu	The text is rendered over the icons
2016-07-21 16:55:41	SylvieLorxu	lol
2016-07-21 16:55:47	@kido	... :)
2016-07-21 16:55:55	@kido	still the same one to blame
2016-07-21 16:56:25	SylvieLorxu	Oh wait I didn't remove everything in GridItem correctly
2016-07-21 16:56:44	@kido	else this is it https://github.com/AsteroidOS/asteroid-settings/commit/4fe05f81d780066891d1e85796c57a95ee83c9e4
2016-07-21 16:57:21	SylvieLorxu	Nah, it was me forgetting to remove everything
2016-07-21 16:58:29	SylvieLorxu	https://lut.im/e4U4j4jWxK/4BBedqcijCJ00tHP.jpg
2016-07-21 16:59:14	SylvieLorxu	Frankly, I find this to look just as well, although I'd like the current choice to be bigger and stuff but... we can do that later I think?
2016-07-21 16:59:22	SylvieLorxu	I would personally say "Let's just use this everywhere"
2016-07-21 16:59:42	SylvieLorxu	Device-specific differences are, after all, a hell to maintain
2016-07-21 16:59:53	@kido	what the heck is wrong with the font
2016-07-21 17:00:08	@kido	oh my, me again
2016-07-21 17:00:15	SylvieLorxu	lol, the font recently changed
2016-07-21 17:00:17	SylvieLorxu	:P
2016-07-21 17:00:52	@kido	I moved the fonts to /usr/lib/fonts here https://github.com/AsteroidOS/meta-asteroid/commit/57476b8010573463a031aa0037b8a24d680a05be
2016-07-21 17:01:07	@kido	which seemed to work fine
2016-07-21 17:01:25	SylvieLorxu	Between the last time I built and now the font changed though
2016-07-21 17:01:26	@kido	but I recompiled everything and I noticed that it only worked if I would put the fonts to /usr/share/fonts
2016-07-21 17:01:31	@kido	so I changed it again 3 days ago https://github.com/AsteroidOS/meta-asteroid/commit/c26c5f8021b45f158df57522770f915b97c67bd3
2016-07-21 17:01:49	SylvieLorxu	Not sure if I have that in my build
2016-07-21 17:01:59	@kido	check what directory they are in and what directory they should be in
2016-07-21 17:02:15	SylvieLorxu	Ah yeah I have the /usr/share/fonts in that file
2016-07-21 17:02:18	SylvieLorxu	Let me check where they are
2016-07-21 17:03:08	@kido	but I changed it back for a good reason... maybe I was working on a different version of Qt or whatever
2016-07-21 17:03:18	SylvieLorxu	/usr/lib/fonts and /usr/share/fonts both contains fonts but share only some OpenSans things and a "truetype" file and /usr/lib/fonts a whole lot of them
2016-07-21 17:03:20	@kido	I have so many build trees atm
2016-07-21 17:03:32	@kido	yes, I'm talking about OpenSans
2016-07-21 17:04:02	SylvieLorxu	Well, those are in /usr/share/fonts, yes
2016-07-21 17:04:43	@kido	then you can try to move them to /usr/lib/fonts but I gotta find out why it didn't work on my watch
2016-07-21 17:05:45	SylvieLorxu	Wow, that's a hell of a difference lol
2016-07-21 17:07:32	SylvieLorxu	By the way, after changing language, you need to reboot (or run "adb shell systemctl restart user@1000"). I could set the language thing up to automatically reboot after that but as that would most likely make us end up in Android Wear... I suppose I shouldn't?
2016-07-21 17:08:09	SylvieLorxu	It's a developer preview, I suppose we can get away with telling people to manually restart? :P
2016-07-21 17:08:19	SylvieLorxu	Unless we can run that systemctl command through DBUS somehow
2016-07-21 17:08:21	@kido	I suppose you don't necessarily need to reboot but just restart the user session
2016-07-21 17:08:31	SylvieLorxu	I know you have DBUS possibility in QML
2016-07-21 17:08:33	SylvieLorxu	Yeah
2016-07-21 17:08:54	SylvieLorxu	Well, actually, we just need to have systemd read that environment variable again
2016-07-21 17:09:04	SylvieLorxu	Actually, no, restart user session because watchface
2016-07-21 17:09:24	SylvieLorxu	Running "systemctl restart user@1000" would work great
2016-07-21 17:09:38	@kido	best solution would be to restart the session through dbus from nemo-qml-plugin-systemsettings
2016-07-21 17:09:55	@kido	else, rebooting (even to android) is ok
2016-07-21 17:10:08	SylvieLorxu	Yeah, I found systemd's dbus thing but... I have no experience with dbus and it's REALLY confusing: https://www.freedesktop.org/wiki/Software/systemd/dbus/
2016-07-21 17:10:44	SylvieLorxu	I saw you used dbus in some other pages so I figured I'd ask
2016-07-21 17:10:50	@kido	https://git.merproject.org/mer-core/nemo-qml-plugin-systemsettings/blob/master/src/languagemodel.cpp#L204
2016-07-21 17:11:16	@kido	this calls the req_reboot methode of dsme
2016-07-21 17:11:26	@kido	method*
2016-07-21 17:11:30	SylvieLorxu	Yeah, but that's a complete device reboot
2016-07-21 17:11:45	@kido	change the service to whatever systemd wants
2016-07-21 17:12:18	@kido	RestartUnit
2016-07-21 17:12:41	@kido	name is probably "user@1000"
2016-07-21 17:13:06	@kido	interface is /org/freedesktop/systemd1 
2016-07-21 17:13:24	@kido	well no, interface is org.freedesktop.systemd1.Manager
2016-07-21 17:13:46	@kido	but service is /org/freedesktop/systemd1
2016-07-21 17:14:45	@kido	I suppose that QDBusInterface systemdInterface("org.freedesktop.systemd1", "/org/freedesktop/systemd1/Manager", "org.freedesktop.systemd1.Manager", QDBusConnection::systemBus()); would work
2016-07-21 17:15:07	SylvieLorxu	I was actually thinking of doing it in LanguagePage.qml, like you do on USBPage.qml
2016-07-21 17:15:13	SylvieLorxu	Seemed more consistent
2016-07-21 17:15:19	@kido	and then call("req_reboot");
2016-07-21 17:15:22	@kido	nah
2016-07-21 17:15:48	@kido	call("RestartUnit", "user@1000", "whatever mode is??");
2016-07-21 17:16:30	@kido	you might need special rights set to the ceres user though
2016-07-21 17:16:35	@kido	in the dbus config
2016-07-21 17:17:18	@kido	and mmhh I don't really mind, I thought that updateMode = UpdateAndReboot would get the job done in a clear way
2016-07-21 17:17:36	@kido	choose what suits you the best as long as the code is well written
2016-07-21 17:19:29	SylvieLorxu	On that note, I cheated because I couldn't figure out how to access that Enum there, so I just pass a number ^^"
2016-07-21 17:19:32	SylvieLorxu	Not clean, I know
2016-07-21 17:20:16	@kido	we will check that later
2016-07-21 17:20:20	@kido	but that needs to be changed
2016-07-21 17:20:32	SylvieLorxu	Definitely
2016-07-21 17:34:56	SylvieLorxu	Nokius: You were the one with a jolla phone, right? Could you give me the /usr/share/jolla-supported-languages/fr.conf file? The proper spelling of French contains special characters so I want to make sure that we do that right
2016-07-21 17:37:41	Nokius	sec
2016-07-21 17:39:52	Nokius	SylvieLorxu: https://gist.github.com/Nokius/bbd6aeee87a13252ce8b619a9f4f7947
2016-07-21 17:40:09	SylvieLorxu	Nokius: Thanks!
2016-07-21 17:41:13	Nokius	SylvieLorxu: are Berlin based kido toled me about someone with S in nick  it's about the Qt days Asteroid talk
2016-07-21 17:41:28	Nokius	for hackers and non attendants
2016-07-21 17:41:46	SylvieLorxu	Nokius: Nope, I live in the Netherlands
2016-07-21 17:42:55	Nokius	SylvieLorxu: nice place hehe
2016-07-21 17:42:57	@kido	Nokius: I was talking about bebecede
2016-07-21 17:43:02	SylvieLorxu	Haha
2016-07-21 17:43:21	SylvieLorxu	I guess "c" sounds like "s" under some conditions :P
2016-07-21 17:43:21	Nokius	kido: didn't u say S...
2016-07-21 17:43:34	 *	Nokius getting old and having to much stuff in my head
2016-07-21 17:43:45	@kido	Nokius: I'm pretty sure I didn't :)
2016-07-21 17:44:17	Nokius	:)
2016-07-21 17:55:01	SylvieLorxu	Pff, guess I should build AsteroidOS one last time just to make sure the build scripts are completely correct
2016-07-21 17:55:08	SylvieLorxu	What a hassle :P
2016-07-21 18:04:43	@kido	SylvieLorxu: it will only rebuild what's needed
2016-07-21 18:04:50	@kido	so it won't take long
2016-07-21 18:04:55	@kido	hi Lrs121
2016-07-21 18:05:12	Lrs121	hello
2016-07-21 18:05:16	@kido	do you have gdb and valgrind with you ?
2016-07-21 18:06:08	@kido	we need the traces from strace, gdb and valgrind and if we can't find it ourselves we will ask the help of our sailfishos porters friends :)
2016-07-21 18:11:40	SylvieLorxu	ERROR: ExpansionError during parsing /home/sylvie/Programs/Forks/asteroid/src/meta-asteroid/recipes-asteroid/asteroid-launcher/asteroid-launcher_git.bb: Failure expanding variable do_fetch[file-checksums], expression was ${@bb.fetch.get_checksum_file_list(d)}  ${@get_lic_checksum_file_list(d)} which triggered exception MalformedUrl: The URL: '' is invalid and cannot be interpreted
2016-07-21 18:11:45	SylvieLorxu	Okay, that... is confusing
2016-07-21 18:12:14	SylvieLorxu	Oh nevermind I see it already
2016-07-21 18:14:14	Lrs121	yah, let me pull up an linux environment
2016-07-21 18:16:46	@kido	:)
2016-07-21 19:13:00	Lrs121	kido: you werent kidding when you said the image would be larger
2016-07-21 19:13:16	@kido	ahah, no I'm never kidding
2016-07-21 19:13:54	@kido	well you get the debugging symbols of asteroid-launcher and all of its dependencies (for example the whole qt debug packages)
2016-07-21 19:14:11	@kido	+ the source code included etc
2016-07-21 19:14:22	@kido	it wheight a lot but makes debugging so much easier
2016-07-21 19:23:38	Lrs121	yup, it pushes the limit of available free space of the watch
2016-07-21 20:12:41	AndroUser	I think my irc app has a problem with multiple wlan access points...  I hate it...
2016-07-21 20:12:50	AndroUser	*itsme
2016-07-21 20:15:35	@kido	itsme____: seems likely
2016-07-21 20:15:38	@kido	wb TheAppleMan 
2016-07-21 21:23:44	Lrs121	kido: after a lot of trickery, ive got the new image on the watch
2016-07-21 21:24:02	@kido	nice :) you hadn't enough space?
2016-07-21 21:24:17	@kido	usualy a debug image fits in my watch
2016-07-21 21:25:24	Lrs121	it fits, but i had to do a bunch of stuff to work arround the limited space on the linux live boot im using
2016-07-21 21:26:17	@kido	ok
2016-07-21 21:27:09	Lrs121	what is the first step now?
2016-07-21 21:28:44	@kido	then start
2016-07-21 21:29:02	@kido	until you see a SIGSEGV or SIGFPE you can usualy continue with c
2016-07-21 21:29:24	@kido	whenever the app fully crashes use "bt" and paste me everything
2016-07-21 21:34:29	Lrs121	this is the only thing i get. http://pastebin.com/gJCiSFFK
2016-07-21 21:35:28	@kido	fuck it segfaults before loading the binary
2016-07-21 21:35:44	@kido	I had the same behavior, no way to know how to debug that though
2016-07-21 21:35:53	@kido	well, gdb won't be of any help then
2016-07-21 21:40:03	Lrs121	bugger, and one valgrind log. http://pastebin.com/t5NXDMGh
2016-07-21 21:41:58	@kido	ok, so that's not very helpful either
2016-07-21 21:43:22	@kido	https://www.google.fr/search?q=site%3Amerproject.org+Floating+point+exception+porters and there seems to be nothing of interest in the irc logs of sailfishos-porters
2016-07-21 21:47:40	@kido	mhhh but I want to know where this sigfpe comes from
2016-07-21 21:54:39	Lrs121	here you are. http://pastebin.com/SJE0WpRA
2016-07-21 21:55:01	Lrs121	ive gotta go get lunch i should be back in ~20 minutes
2016-07-21 21:56:20	@kido	ok
2016-07-21 21:57:39	@kido	so it's not helpful either
2016-07-21 22:00:19	@kido	could you send me the output of pstree? just to check that android services are running, and then try to use EGL_PLATFORM=hwcomposer test_hwcomposer (also with strace, gdb and valgrind)
2016-07-21 22:01:10	@kido	also, remove /vendor/lib/egl/libGLES.so before
2016-07-21 22:01:36	@kido	once you've done that, you can re-try everything with EGL_PLATFORM=fbdev instead of hwcomposer
2016-07-21 22:17:22	Lrs121	heres pstree. http://pastebin.com/Mem3i7u0
2016-07-21 22:21:29	Lrs121	strace test_hwcomposer. http://pastebin.com/JazJPGZY
2016-07-21 22:23:57	Lrs121	gdb test_hwcomposer. http://pastebin.com/anvWpLDD
2016-07-21 22:25:49	Lrs121	valgrind test_hwcomposer
2016-07-21 22:25:52	Lrs121	http://pastebin.com/jLBP78jH
2016-07-21 22:29:00	Lrs121	kido: with fbdev am i still passing test_hwcomposer?
2016-07-21 22:29:04	@kido	yep
2016-07-21 22:29:26	@kido	with fbdev, strace should be enough
2016-07-21 22:29:32	@kido	(no gdb or valgrind)
2016-07-21 22:31:29	Lrs121	ok, heres the strace fbdev. http://pastebin.com/vHmZ9nEh
2016-07-21 22:41:57	@kido	Lrs121: do you still have your android build directory ?
2016-07-21 22:43:37	@kido	also, what android wear version did you pull ? (you should be able to find that in /system/build.prop)
2016-07-21 22:49:12	Lrs121	i've got a new build directory for cleaning up the manifest. it's version 6.0.1 build MEC23U.
2016-07-21 22:50:13	@kido	well, first things first, all ports so far have been made on top of 5.1.1
2016-07-21 22:50:20	Lrs121	the zenwatch 2 shipped with lollipop but quickly updated. mine shipped with 6.0.1 since i got it after the update was sent
2016-07-21 22:51:09	@kido	and I think you should do the same, I'm not sure libhybris works correctly with marshmallow (maybe I'm wrong but let's get rid of uncertainties)
2016-07-21 22:51:31	@kido	and the android parts you built are designed for 5.1.1 so there is probably an issue there
2016-07-21 22:52:00	@kido	we need to find a backport of 5.1.1 on xda
2016-07-21 22:52:23	Lrs121	that should be relatively easy.
2016-07-21 22:58:37	@kido	https://drive.google.com/folderview?id=0B1-WblgnnTdZVWk2bXNnRUpSUGc&tid=0B1-WblgnnTdZZHhTVE9XY0hPSjg according to recovery.log it seems to be 5.1.1
2016-07-21 22:59:21	@kido	so I suppose you should be able to extract a good system directory from it
2016-07-21 23:00:22	@kido	pull /system/, keep it somewhere warm in your hdd and we will subject it to any kind of experiment
2016-07-21 23:00:54	@kido	start by re-following the instructions from README-system-dir (no need to rebuild everything, of course)
2016-07-21 23:01:48	@kido	if this works alright then we're fine, else we are gonna try to replace the least amount of files possible (e.g: just /system/lib/libc.so and /system/lib/libdsyscalls.so)
2016-07-21 23:02:28	@kido	but that's probably not enough since we need to rebuild a couple of things with QCOM_BSP ... but anyway, let's see
2016-07-21 23:02:53	Lrs121	we cant use that one. the build id on the tail end of the directory name is the same as what i based my port on
2016-07-21 23:06:22	@kido	mhh
2016-07-21 23:07:39	@kido	right, this seems to be marshmallow indeed
2016-07-21 23:07:52	@kido	I don't know where we could find an older build though
2016-07-21 23:09:14	Lrs121	i found one, but i believe its 5.0.1
2016-07-21 23:10:05	@kido	http://www.hanfei.name/ota-zip-asus-zenwatch-android-wear-5-1-1-lollipop-update-rolling.html
2016-07-21 23:10:58	@kido	mhh it's zenwatch 1 sorry
2016-07-21 23:16:29	@kido	I can't find LCB51
2016-07-21 23:17:40	@kido	maybe you can flash 5.0.1 and get OTA from there until you reach 5.1.1
2016-07-21 23:17:53	@kido	quite painful though...
2016-07-21 23:19:30	Lrs121	that it would, i can attempt that as a last option since i have a backup of the latest.
2016-07-21 23:20:00	@kido	yep
2016-07-21 23:20:20	Lrs121	what concerns me is the bootloader which is the latest but may not play nice with the update
2016-07-21 23:20:22	@kido	you can also try to hack on a 5.0.1 /system folder but i'm not sure it would work either
2016-07-21 23:20:56	@kido	mmhhh... I don't know, this seems weird to me that a bootloader doesn't work with a newer kernel but it's true that I had an issue like that with the sony sw3
2016-07-21 23:21:50	Lrs121	im more worried that there will be a check for a certain bootloader version like there is on the nexus devices
2016-07-21 23:22:23	@kido	never heard of that before
2016-07-21 23:43:50	Lrs121	ill take a look and see what happens when i get home. i dont entirely trust the linux live environment ive got
2016-07-21 23:45:16	@kido	ok, I'm off