Tuesday, 2020-10-06

T42<edp_17> @elros34: I created a fresh build. The audio is fixed with the xx.pa and custom audio_policy.conf. However, none of the sensors are working (including gps). The user nemo was in the system group but even if I removed it and rebooted, the gps didn't start working. Plus I have no idea why none of the sensors are working. Is it a permissi06:30
T42on issue? What should I check?06:30
T42<edp_17> Of course I have added that '%define additional_ha_groups \ system\ %{nil}' to your droid-hal-$DEVICE.spec and rebuilt droid-hal and all the packages too.06:31
T42<edp_17> And just realised that bluetooth also stopped working. 😭 All these worked on previous 3.3.0.16 builds, what did happen?07:41
T42<edp_17> Should I try to build 3.4? Will that solve these?07:41
T42<elros34> @edp_17 I told about sensors that you need to downgrade sensorfw plugin (hybris-libsensorfw-qt5-hal). For 3.4.0.x use 0.11.9, for 3.3.0.x check on device which version of sensorfw-qt5 you have installed and use same version for plugin10:19
T42<elros34> Did you remove that directory /data/system/gps before first boot?10:21
T42<elros34> or even now, it doesn't matter. Just get rid of previously created files. I assume the issue is same: gps crash10:35
T42<elros34> gpsd*10:35
*** Sellerie0 is now known as Sellerie14:44
T42<edp_17> So, the issue with the GPS is not the user's permission but this file '/data/system/gps/.gps.interface.pipe.to_jni'?15:25
T42<edp_17> I'll try the rest. Thank you!15:26
T42<elros34> I don't know, you never show ls -al of that directory but can imagine that it has now wrong permission from previous test15:27
T42<edp_17> You never asked. otherwise I would show that. :P15:32
T42<edp_17> I'll get rid of all files in that folder and retry.15:32
T42<edp_17> Thanks.15:32
T42<elros34> @edp_17 btw I've checked your wacom driver and it might work with evdevmouse too. Add '-plugin evdevmouse:abs:/dev/input/event3' and I think cursor should be visible. For button support small change would be needed16:54
T42<edp_17> @elros34: Sounds brilliant! Thanks. I'll try this.17:39
T42<edp_17> About the gps issue. Interesting because it doesn't work right after the first boot on a freshly installed build. At that point there is no file in that folder from a previous attempt.17:41
piggzrinigus: you learn something new every day ... didnt realise you can upload multiple files at once on openrepos18:17
piggzatleast makes uploading my kf5 files tolerable18:17
riniguspiggz: you can :) . have used that to upload language support packages together for maliit predictor.18:19
piggzmal: can i make a feature request please.....18:19
piggzDisplay Settings > Keep screen on in developer mode18:20
piggzmakes much more sense than keep screen on while charging!18:20
malpiggz: hmm, need to think about that18:34
attahI like keep screen on while charging... for example when monitoring an update, but i'm also all for more granular control19:47
piggzattah: but you wouldnt want it on all night if charging20:09
piggzright?20:09
attahWell, i usually manually lock the display... so not really an issue for me.20:10
attahThat still shuts it off just fine...20:10
attahI can go weeks or months with the setting on and not even notice. (just did as a matter of fact)20:12
T42<edp_17> @elros34 : Are you around?20:52
T42<elros34> just ask20:53
T42<edp_17> About the GPS (still)20:54
T42<edp_17> After the reboot nemo always has the group system. (Even if I do a 'gpasswd -d nemo system' before the reboot)20:55
T42<edp_17> After the reboot the file '/data/system/gps/.gps.interface.pipe.to_jni' is always missing however, it gets generated when I start gpsinfo.20:56
T42<edp_17> GPS doesn't work with this setup.20:56
T42<edp_17> However if I chmod a+rw that file and restart gpsinfo, the location icon start flashing at the top.20:56
T42<edp_17> My questions: 1. should nemo has the group system or not (or it doesn't matter)?20:57
T42<elros34> it doesn't harm if nemo i in system group. It can cause some issues if it's not20:58
T42<edp_17> 2. how can I permanently change the permission of the file '/data/system/gps/.gps.interface.pipe.to_jni'? Or how can I do the chmod a+rw before the geoclue-hybris starts or after the creation?20:58
T42<edp_17> Ok, then I leave nemo in that group.20:59
T42<edp_17> Wow, now I rebooted again and I got the following:21:00
T42<edp_17> Nemo was in system group, file '/data/system/gps/.gps.interface.pipe.to_jni' was missing. I toggled the location on and started the gpsinfo app and the file '/data/system/gps/.gps.interface.pipe.to_jni' has recreated but with a different permission than before!21:01
T42<edp_17> Before it was 'prw-rw----    1 root      system' and now 'prw-rw----    1 gps      system'21:02
T42<edp_17> And the location icon started to flash.21:02
T42<edp_17> I don't understand what is going on and why. 😱21:02
T42<edp_17> The GPS is working now but I have a feeling if I flash this build again, it won't start again.21:03
T42<elros34> have you removed /data/system/gps before all of your tests not just that file? This directory should be recreated with correct permissions somewhere in your /init*.rc21:07
T42<elros34> even with root:system gpsd should be fine. You could strace it21:10
T42<edp_17> No, I have not done that. Honestly I never removed the /data/system/gps21:10
T42<edp_17> If I rebuild sensorfw, what .spec I should use sensorfw-qt5.spec or sensorfw-qt5-hybris.spec?21:13
T42<elros34> sensorfw-qt5 is built and downloaded from jolla repos so you only need hybris plugin21:14
T42<edp_17> Okay. What happens if I build sensorfw-qt5.spec?21:16
T42<edp_17> I think what I said is not clear. Sorry, I explain.21:19
T42<edp_17> So, you said to downgrade sensorfd to 0.11.9. Therefore I found the commit to 0.11.9 in sensorfw (89ed1ef63e080b1097477f4a687201ce3683c222) and I fetched then git resetted to that commit my local sensorfw and then I tried to build with 'rpm/dhd/helpers/build_packages.sh -D -o -b hybris/mw/sensorfw -s rpm/sensorfw-qt5.spec'21:20
T42<edp_17> My question about using sensorfw-qt5.spec or sensorfw-qt5-hybris.spec was related to this command.21:21
T42<elros34> sensorfw-qt5-hybris.spec21:28
T42<edp_17> Okay. I hope that won't be a problem if I used too the other .spec. 😊21:32
T42<elros34> 0.11.9 was for 3.4.0.x but it might be ok for 3.3.0.x too. You will see:)21:41
T42<edp_17> 😄 I am installing it now.21:44
T42<edp_17> Yep. Now all sensors are working again. Thanks for your support.21:49
T42<edp_17> Except the GPS. :P21:50
T42<edp_17> Permission of '/data/system/gps/.gps.interface.pipe.to_jni' is again: prw-rw----    1 root     nemo21:50
T42Patrick %lastname% was added by: Patrick %lastname%21:53
T42<edp_17> Yeah, I need to chmod a+rw that file after every reboot once it is created. 😭21:55
T42<edp_17> I only found the /data/system/gps related stuff in one .rc file (init.universal5433.rc). Here is the related stuff (I also included the ls -al output too): https://paste.ubuntu.com/p/m79Spvmckb/22:04
T42<edp_17> @elros34 : Regarding the wacom stuff, I have added that into 'droid-hal-device.conf' and rebooted, but nothing happens when I pull out the s-pen. Is there anything else I need to add/change?22:19
T42<elros34> I thought that should be enough provided you have mousetracker patch enabled and in log there is clean indication that evdevmouse is using that device22:22
T42<edp_17> mousetracker is not installed (it was but not on this image). I am installing it now.22:23
T42<edp_17> You were right. With mousetracker installed the cursor is visible! >:D22:27
T42<edp_17> What do I need to do to make it work?22:27
T42<elros34> replace BTN_TOUCH with BTN_LEFT in all files in wacom/ directory22:29
T42<edp_17> Done. There were two files (wacom_i2c.c and wacom_i2c_func.c)22:33
T42<edp_17> I guess now I need to rebuild hybris-hal and what else?22:34
T42<elros34> hybris-hal builds everything22:40
T42<edp_17> Okay, that's also done. :D22:41
T42<edp_17> New hybris-boot.img is installed ...... and it works!!!22:46
T42<edp_17> Cool! 😁 Is there a way to emulate right click?22:47
T42<elros34> If you have some other button on stylus then maybe assign BTN_RIGHT to it but I guess sfos doesn't support it in any qml components22:58
T42<edp_17> Yes, there is another button on its side.22:59
T42<edp_17> However, the evdev_trace doesn't report anything when I press that.23:00
T42<edp_17> I guess if I figured out which button is the side one and assign the BTN_RIGHT to it, then I will be able to test it in a Desktop Linux within the LXC. What do you think?23:03
T42<elros34> That would be strange if evdev_trace -t doesn't print anything.23:05
T42<elros34> unless that button is used switch some mode but in driver there are few buttons defined23:08
T42<edp_17> I have tested the button. Here I pulled out the pen, pressed the side button, then touched the screen and pressed the side button few times and slide the pen back: https://paste.ubuntu.com/p/vT77tWyH4X/23:13
T42<edp_17> Then the same again but without pressing the side button this time: https://paste.ubuntu.com/p/9ZXGkNWXtQ/23:13
T42<edp_17> The difference is: BTN_STYLUS, so I think that is the side button.23:13
T42<Patrick %lastname%> I'm looking for someone in Rome, Italy, good at porting Sailfish OS to Esperia phones. Anyone?23:16
T42<Patrick %lastname%> Sorry, I obviously meant Xperia phones23:30

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!