Monday, 2019-09-02

UmeaboyGotta go. It's very late here. Take care.00:07
spiiroinkimmoli rinigus: that act dead thing: act dead ui uses statefs for battery/charger info, shutdown logic etc (in dsme) uses mce. in most of the devices, most of the time these two agree. when either has problems, it is going to be different kind of problems04:16
spiiroinfor example for new ports you might need to blacklist power supply udev devices if they are causing problems for mce side logic04:19
spiiroinsome info at:
kimmolispiiroin: ok, not a new port. it used to work on my device, before 304:43
kimmolibefore 3.0.x04:43
spiiroinkimmoli: the statefs deprecation started at mce 1.96.0 / around 3.0.[01], quite a bit of tweaking in mce 1.99.2 / 3.0.3 -> at which state all officially supported devices were ok afaik05:44
spiiroinbut given how much variance there is in these charger/battery things between various devices, I would not be surprised if something that has not been explicitly checked might have problems05:45
kimmolispiiroin: ok. interesting. more interesting is that it works on rinigus device, but not on my..05:52
riniguskimmoli, spiiroin: works is a touch strong word. It did shutdown after 100% on charger and when I pulled power out. However, later boot failed to get into gui. Had to power off and boot again. Still have to check the logs and see what's going on, journal should be there05:57
spiiroinkimmoli, rinigus: there were things like: some power_supply device shows up in udev, but never sends change notifications -> if it was used as false positive "is a charger" caused issues that varied depending on whether charger was connected during bootup or not06:09
spiiroinwhat device you're having these problems on? qc? mtk? something else?06:10
spiiroinquick test: ssh in via wlan, run mce manually: "mce -Tq -l modules/battery-udev.c:*" play with various chargers, see if the udev events vs mcebat_update stuff makes sense06:12
kimmolionyx, qc06:12
kimmoliit detects the charger disconnect, at least UI shows so06:13
spiiroinkimmoli: a rough divide: qc devices have odd device nodes, mtk devices have funky battery full behavior06:13
kimmoliyeah this has also something in kernel bms with the battery %06:14
kimmoliit might show 100% for half a day, or loop between 95-100%, then drop to 50%06:14
kimmolithe driver has some stupid smoothing for-loop-algorithm06:15
spiiroinkimmoli: what that (act dead) ui shows kind of does not matter. now it uses statefs context props, which will be cleaned off, act dead ui probably among the last06:20
spiiroinkimmoli rinigus: fyi, for example, some recent changes:
electro575hi, how it's possible to install mail client android please ?06:51
electro575@elros34 : what do you use as client email ? is it possible to install android app ?09:00
electro575thank you09:00
T42<elros34> I use default email client from jolla. Only official devices support android applications (alien dalvik).09:02
electro575humm okey09:12
T42<DylanVanAssche> Hi :)09:45
T42<DylanVanAssche> I looked into the systemd service files for `geoclue`, however I can't find out how it's started?09:45
T42<DylanVanAssche> I would like to run it manually, but it's a non existing command `geoclue`. Geoclue is installed though...09:45
T42<DylanVanAssche> Maybe I'm just looking over it? Upstream docs are not useful since our geoclue is ancient09:45
Mister_Magister@DylanVanAssche check source09:54
T42<adampigg> geoclue is a dbus service10:23
T42<adampigg> rpm -ql geoclue helps10:23
T42<DylanVanAssche> @Mister_Magister I checked the source but I couldn't figure it out ;p10:30
T42<DylanVanAssche> @adampigg That's useful :D I didn't know that command :O10:31
T42<DylanVanAssche> In this case, is it possible to not run it as a service? For debugging reasons10:32
riniguskimmoli and spiiroin: pasted 3 logs recorded consequently during actdead -> failed boot -> successful boot . paste at
rinigus^ on onyx
T42<adampigg> dylan: try running the exe manually ... presumably it will register on the bus and work10:56
T42<DylanVanAssche> @adampigg thanks! It works :D11:24
T42<DylanVanAssche> (Photo, 1451x2560)
Mister_Magistertat reminds me @adampigg we need to fix it on pro111:28
T42<adampigg> we do, but we both know the state of my /pro 1 :(11:29
* Mister_Magister *wink wink*11:29
Mister_Magister@adampigg i don't want to advertise myself but are you using microtube yet? :P i remember i added your game preinstalled on my ports back in the day xd (not anymoar tho)11:30
T42<adampigg> im not ... still LLS, becuase u dont sync my subscriptions to my YT account :P11:30
Mister_Magister@adampigg stay tuned for 2.0 ;)11:31
T42<adampigg> i heard u were adding it11:31
electro575@elros : what do you use for encrypt your email as PGP ?12:05
T42<elros34> I don't encrypt my emails :)12:06
T42<DylanVanAssche> Hi :)16:31
T42<DylanVanAssche> We have a modem with GPS on the PinePhone, to enable GPS/NMEA messages we need to write an AT command to the AT interface of the modem.16:31
T42<DylanVanAssche> Currently, we have a systemd service which enables the GPS interface at boot, in terms of power consumption, we should only enable the GPS when location is enabled on the phone (and disable it if requested).16:31
T42<DylanVanAssche> Is there any trigger or interface we can use for this purpose?16:31
T42<adampigg> I assume that the ui sends a dbus message to geoclue, so we might need to find a way to trigger an action when that happens16:35
T42<DylanVanAssche> I would suspect the same... I just tested it, when location is turned off in the UI, it still works :D16:36
T42<adampigg> gpsd device-hook to the rescue17:05
T42<DylanVanAssche> :D17:23
r0kk3rzsounds like it needs a proper driver17:32
T42<adampigg> R0kk3rz, what do u mean18:22
T42<DylanVanAssche> @r0kk3rz Geoclue is not respecting the location enable/disable button in the UI. gpsd enables and disables the GPS hardware on the client's request now. In this situation, geoclue is a client of gpsd. So if location is disabled in the UI, geoclue should not contact gpsd. The app that request a location should get an invalid/las18:24
T42t location from geoclue18:24
r0kk3rzgeoclue is pretty old and thr maintainer wont care18:26
T42<adampigg> On the plus side, better to work too much, than not at all!18:26
T42<adampigg> Yeah,  i dont understand why our version is soo old...if we had a new one, we could share location over network to non gps tablet like latte18:27
T42<DylanVanAssche> Indeed! It's so ancient...18:29
T42<DylanVanAssche> Maybe it's related to the hybris provider plugins?18:29
T42<DylanVanAssche> Just not maintained then?18:34
r0kk3rzmostly that geoclue2 is a complete rewrite and incompatible18:34
r0kk3rzso updating is a lot of work18:34
T42<DylanVanAssche> Oh that's indeed hell18:34
T42<DylanVanAssche> However, we will reach a point in time that we can't avoid it probably18:35
rinigusgeoclue2 is supported by qt only starting from 5.12. so, don't expect it too soon20:52
rinigus... unless I am missing some trick allowing to use geoclue2 in earlier qt versions20:52
r0kk3rzmight be backportable20:53
r0kk3rzmaybe to qt5.920:53
malrinigus: isn't geoclue2 quite limited compared to old geoclue, I couldn't find way to handle satellite information for example, I only saw location20:53
rinigusmal: maybe. I was mainly interested on how to get location on qt5.11 with desktop geoclue2 and found out that I had to bump qt to 5.12.20:55
rinigusas for geoclue2 limitations, I have no idea20:55

Generated by 2.17.1 by Marius Gedminas - find it at!