| spiiroin | if mce can find evdev sources for light / proximity sensors, those are preferred over sensorfwd (more deterministic suspend behavior etc) -> sensorfwd is used for sensor power control, but data sensorfwd sends is ignored | 04:26 |
|---|---|---|
| spiiroin | now, if there is false positive proximity sensor found during evdev probe -> things go wrong | 04:27 |
| spiiroin | this lists what mce thinks various evdev inputs are: mce -Tq --auto-exit=0 -levent-input.c:evin_iomon_device_add | 04:28 |
| spiiroin | (need to stop mce service first) | 04:28 |
| spiiroin | ^ @elros34, @Anton %lastname% (that nick, eh?) | 04:31 |
| rinigus | @erfanoabdi: to my understanding, you need 0.12.1 version of sensorfw for using rotation vector-based approach. or patch your version accordingly. | 05:59 |
| rinigus | I am using sensorfw with hybris, not binder. as such, I don't know if anyone has tested binder-based code paths and whether `d->degrees_ = data.u.vec3.x` would be correct there. | 06:01 |
| rinigus | with hybris, on sony tama, all works very well. | 06:01 |
| rinigus | to use that adaptor (or it's geo variant), you just need to plug it in sensorfw configs, as in https://github.com/sailfishos-sony-tama/droid-config-sony-tama-pie/blob/master/sparse/etc/sensorfw/primaryuse.conf | 06:03 |
| rinigus | as far as I remember and after a glance in the source, compass sensor feeds on compass chain which, in turn, gets data from orientation sensor. see https://git.sailfishos.org/mer-core/sensorfw/blob/master/chains/compasschain/compasschain.cpp#L45 | 06:06 |
| rinigus | and below | 06:06 |
| rinigus | that allows you to use sensorfw.compass in qtsensors: https://github.com/sailfishos-sony-tama/droid-config-sony-tama-pie/blob/master/sparse/etc/xdg/QtProject/Sensors.conf | 06:07 |
| rinigus | confusingly enough, orientationsensor was something else. but I never looked into it, just was guided by helpful folks over here when working on compass | 06:08 |
| rinigus | if your device has broken rotation vector implementation, according to mal as on xperia 10, you could use generic compass from the PR linked by mal last night. that works as well. difference is on whether you calculate it yourself (as in generic) from underlying data or prefer to use orientation vector provided by android. in principle, the latter should be optimized by manufacturer | 06:11 |
| rinigus | mal: why do you think that compass sensor is broken in sensorfw? | 06:13 |
| mal | rinigus: the virtual compass sensor which is not used by anything, at least last time I tried it I couldn't get it to work properly | 08:23 |
| rinigus | mal: is it sensor or adaptor? in sensorfw-speak? | 08:34 |
| mal | rinigus: if you look at https://git.sailfishos.org/mer-core/sensorfw/blob/master/chains/compasschain/compasschain.cpp you can see that if there is no orientation adaptor define then compass is generated automatically | 08:42 |
| rinigus | mal: that's right, understood. | 08:43 |
| Mister_Magister | jusa: somehow i cannot make pa playback at more than 48kHz even though my phone dac should go up to 192kHz. i've added avoid-resampling and changed alternative sampling rate to 96kHz and no luck. Is it due to some pa-module-droid limitation? | 11:27 |
| Mister_Magister | also plz start building pa with soxr kthx | 11:27 |
| Mister_Magister | soxr has no deps so it doesn't cost you anything to build with soxr (i just did it | 11:38 |
| T42 | <edp_17> Hi all, Could somebody help me out with this, please? I am trying to build 3.4.0.24 with CM12.1 base for hammerhead (Nexus 5) and I got this when ran the 'rpm/dhd/helpers/build_packages.sh'. It is complaining about 'hammerhead-bluez-configs' that I never heard of before. Thanks. https://paste.ubuntu.com/p/ZWH8sgdsgQ/ | 12:25 |
| T42 | <erfanoabdi> @rinigus [that allows you to use sensorfw.compass in qts …], Thanks it's helping.. | 12:27 |
| T42 | <erfanoabdi> I'm not porting any device, making a program to read some data from sensorfw dbus directly | 12:27 |
| rinigus | @erfanoabdi: well, it still requires a device with it running :) . so, device support is needed | 12:31 |
| rinigus | @erfanoabdi: that's a small daemon that I am using to detect pickup gesture on tama: https://github.com/sailfishos-sony-tama/pickupd/blob/master/src/sensorreader.cpp | 12:33 |
| rinigus | as that gesture is guided via sensorfw, maybe of help for your program | 12:33 |
| T42 | <erfanoabdi> Thanks yeah it helps | 12:35 |
| T42 | <Anton %lastname%> Can anyone guess why some rules in 999-android-system.rules aren't run? I have that at the end of the file: https://pastebin.com/X0zfMP1P, yet the permissions are never set. | 12:47 |
| mal | did you add those manually? | 12:49 |
| mal | you might want to check if some .rc file is setting those to some different value | 12:50 |
| T42 | <Anton %lastname%> I modified device/rootdir/root/ueventd.qcom.rc to get that. I supposed it could be some rc file, but they shouldn't be started with "udevadm trigger", should they? | 12:51 |
| mal | hmm, I'm a bit lost, usually led stuff is in system/core | 12:53 |
| mal | I mean vibrator stuff also | 12:53 |
| mal | https://github.com/mer-hybris/hybris-patches/commits/hybris-16.0 you can see the vibrator permission fix the there | 12:54 |
| T42 | <Anton %lastname%> Yes, I have that on device. /vendor/etc/hw/init/hw/init.mmi.rc, however resets it back to system:system | 12:58 |
| T42 | <Anton %lastname%> idk how honestly | 12:58 |
| T42 | <Anton %lastname%> I don't want to touch vendor partition, because that's ugly imho | 12:59 |
| mal | hmm | 13:01 |
| T42 | <elros34> if init.mmi.rc resets it to system:system then obviously even if udev rule works then permission will be overwritten | 13:02 |
| T42 | <Anton %lastname%> if i set permissions manually, they stay that way until reboot. Udev just doesn't get to that rule is my thought | 13:04 |
| T42 | <elros34> but it doesn't matter whether udev rule works or not if permission will be overwritten few seconds later by init.mmi.rc. | 13:05 |
| T42 | <elros34> You need bind mount that rc file or run some systemd service before ngfd start and after droid-hal-init | 13:06 |
| T42 | <Anton %lastname%> I set the permissions to system:input manually, run "udevadm trigger" and they stay that way | 13:06 |
| T42 | <Anton %lastname%> I guess bind mounting is an option | 13:06 |
| T42 | <elros34> example how it's done for custom ld.config: https://github.com/mer-hybris/droid-config-sony-ganges-pie/commit/5d98fc2cd583860313648eeac9b7a7264a471ef3 | 13:10 |
| T42 | <adampigg> @Anton %lastname% [I guess bind mounting is an option], i do a bit of bind mounting for such things on pro1/volla | 13:24 |
| Mister_Magister | where the hecc is pa-modules-droid configuration | 13:30 |
| Mister_Magister | i need to add more sinks | 13:30 |
| T42 | <Anton %lastname%> Seems like I'm ready for release. Hardware adaptation-wise it's ready for daily use and installation by typical user, I believe. I will compile the final image now and post it to xda and wherever else. How's OBS and store going with community ports? | 15:34 |
| Thaodan | Please set a username its hard to interact you from the IRC side | 15:37 |
| T42 | <zinstack625 %lastname%> Is that better? | 15:38 |
| Thaodan | https://telegram.org/faq#f-was-sind-benutzernamen-wie-bekomme-ich-einen | 15:41 |
| Thaodan | no | 15:41 |
| T42 | Thaodan was added by: Thaodan | 15:41 |
| T42 | <zinstack625> Found it, hopefully that'll do | 15:48 |
| mal | looks better now | 15:49 |
| Thaodan | Yes thank you | 15:50 |
| T42 | <zinstack625> mal, r0kk3rz, sledges, mind giving me OBS and Jolla Store access? | 17:07 |
| piggz_ | mal: what tag of qtbase was used in 3.4 ? | 17:57 |
| Thaodan | piggz_: upgrade-3.4.0 | 17:59 |
| piggz_ | Thaodan: thx | 18:00 |
| Mister_Magister | So how is hybris16 | 20:57 |
| Mister_Magister | 17* | 20:57 |
| Mister_Magister | i need hybris17 | 20:57 |
| T42 | Sara %lastname% was added by: Sara %lastname% | 23:30 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!