| T42 | <edp_17> Somebody told me that in that driver it could be possible to emulate mouse events and after use the stylus as a mouse. | 00:00 |
|---|---|---|
| T42 | <edp_17> But I don't know how to start. | 00:00 |
| T42 | <elros34> quite challenging. If it reports all coords maybe other way would be to handle it in qt evdevmouse plugin? | 00:04 |
| T42 | <edp_17> I have tried that. The evdevmouse plugin is already added into that file and if I attach a mouse, that works but not this s-pen. | 00:05 |
| T42 | <edp_17> What I just spotted my pc (all in one with touch screen) reacts to this pen and I can click on the screen with it. 😜 | 00:06 |
| T42 | <elros34> for sure it would require some modifications. Generating events in kernel are not that hard. Here I generate KEY_NEXTSONG on volume up press and hold | 00:08 |
| T42 | <elros34> https://github.com/elros34/android_kernel_motorola_msm8960-common/blob/672b14020cb5b550e8b1affccbf33d86b5e04fe9/drivers/input/keyboard/pmic8xxx-keypad.c#L360 | 00:08 |
| T42 | <elros34> I don't remember anymore. Have you tried evdevtablet plugin? I think it should be easier to make it work than changing kernel driver. | 00:14 |
| T42 | <edp_17> I don't remember that I tried evdevtablet plugin. But I am doing now. | 00:18 |
| T42 | <edp_17> In the meantime the GPS works. Gpsinfo found some satelite: use 0/7. wooohooo! :) | 00:19 |
| T42 | <edp_17> Does it mean that the GPS works? | 00:19 |
| T42 | <elros34> yes, but 0 in use so there is no fix yet | 00:21 |
| T42 | <edp_17> How can I fix it then? | 00:21 |
| T42 | <elros34> when you are testing evdev plugins add QT_LOGGING_RULES=qt.qpa.*=true do droid-hal-device.conf | 00:22 |
| T42 | <elros34> about fix: you need to wait so there is good signal for 3 or 4 satelites and it should get a fix quite fast then | 00:22 |
| T42 | <edp_17> Oh, I see, so I need to let it at the window longer. Will it fix the GPS for later use? (I mean will it save something into the /etc/gps.conf?) | 00:24 |
| T42 | <edp_17> I have added the plugin and the logging stuff into droid-hal-device.conf. Rebooting. | 00:25 |
| T42 | <edp_17> And it is not working. :) (I expected that.) | 00:28 |
| T42 | <elros34> I doubt anything is saved in /etc/gps.conf. If it get a fix then you will have your position, thats it. | 00:29 |
| T42 | <edp_17> Okay. And in the future I don't need to start the geoclue-hybris with debug, do I? | 00:30 |
| T42 | <elros34> it's started automagically via dbus | 00:31 |
| T42 | <edp_17> Now I rebooted, toggled the Location on and started gpsinfo. The Location icon is not flashing at the top. | 00:31 |
| T42 | <elros34> does gpsd crash? | 00:31 |
| T42 | <edp_17> Seemingly not. (The 'journalctl -f --no-tail | grep gpsd' doesn't show anything but the gpsinfo app is running.) | 00:32 |
| T42 | <edp_17> And it has the same pid. | 00:33 |
| T42 | <elros34> then this might be a bug I wanted to debug tonight:) Try to disable and enable gps in setings then start gpsinfo again, csd | 00:33 |
| T42 | <edp_17> Okay. | 00:35 |
| T42 | <edp_17> The 'High-accuracy' was removed and the custom was selected. I ticked 'High-accuracy' then turned off then on the location and now the gps test in csd immediately found satelites. | 00:37 |
| T42 | <edp_17> I guess the gps works for me on this device because I re-built the geoclue-providers-hybris. Am I right if I think that if I build an sfos image now, that will include this geoclue-providers-hybris-xx.rpm and the gps will work on that build? | 00:40 |
| T42 | <elros34> I think it was because of some leftovers with wrong permission in /data/system/gps/ and/or lack of nemo in system group. | 00:42 |
| T42 | <edp_17> Do you mean on the device? | 00:44 |
| T42 | <edp_17> How can I add nemo into system group in the sdk? | 00:44 |
| T42 | <elros34> remove nemo from system group now: gpasswd -d nemo system. If ater reboot it issue will return then: https://github.com/mer-hybris/droid-hal-device/commit/1dac6f1ff3b58215003af61ba52bd91c5b177219 | 00:48 |
| T42 | <edp_17> I removed nemo from system group, rebooted and gps works. | 00:54 |
| T42 | <edp_17> That commit (https://github.com/mer-hybris/droid-hal-device/commit/1dac6f1ff3b58215003af61ba52bd91c5b177219) is already in my sdk. | 00:56 |
| T42 | <edp_17> Oh, I see now. As the dhd has changed but I used an already built image, the permission/groups were not right on the device. This will not be a problem on a fresh build as it is fixed in droid-hal-device.inc. | 00:58 |
| T42 | <elros34> Not sure if that was the issue, you will have to remove /data/system/gps so it' recreated | 01:02 |
| T42 | <elros34> ant then compare permissions | 01:02 |
| T42 | <elros34> about evdevtablet: did you have 'evdevtablet: using /dev/input/eventx' in logs? | 01:03 |
| T42 | <edp_17> Yeah, I am just trying to record what the solution was and I am a bit puzzled. First I added nemo into system group, then I deleted the /data/system/gps/.gps.interface.pipe.to_jni file and removed nemo from the system group. So basically removing that file has solved the issue. | 01:04 |
| T42 | <edp_17> evdevtablet: what logs are you referring to, journalctl? | 01:05 |
| T42 | <elros34> yes | 01:06 |
| T42 | <edp_17> The 'journalctl -f --no-tail | grep evdevtablet' doesn't show me anything. | 01:06 |
| T42 | <elros34> have you added QT_LOGGING_RULES=qt.qpa.*=true? it should be more verbose somewhere close to lipstick start logs | 01:07 |
| T42 | <edp_17> Yes I did. Now I added the =/dev/input/event3 after evdevtablet. | 01:08 |
| T42 | <edp_17> And I got: Oct 03 02:08:29 GalaxyNote4 lipstick[4225]: [W] unknown:0 - No such plugin for spec "evdevtablet=/dev/input/event3" | 01:09 |
| T42 | <elros34> I think it should be -plugin evdevtablet:/dev/input/ eventx | 01:12 |
| T42 | <elros34> : instead = | 01:12 |
| T42 | <edp_17> Okay, I have removed the /dev/input/event3 and now I have evdevtablet: using /dev/input/event3 and some other lines in the journalctl. | 01:14 |
| T42 | <edp_17> I have this: https://paste.ubuntu.com/p/NXv2Vqv48F/ | 01:15 |
| T42 | <elros34> looks good, if this device generate ABS_X/Y, BUTTON_TOUCH and EV_SYN then then it should be forwarded further | 01:19 |
| T42 | <edp_17> With edvev_trace, I get: SW_0013 - 1 (pull ot of the pen), SW_0013 - 0 (put back in), when I touch the screen: ABS_X - 4938 and ABS_Y - 6289 and ABS_DISTANCE - 0, and when I move the pen towards the screen: ABS_X - 4938 and ABS_Y - 6289 and ABS_DISTANCE - 177 (the distance is lowering as the pen gets closer to the screen) | 01:22 |
| T42 | <edp_17> How can I forward these events and where to? | 01:23 |
| T42 | <edp_17> Well, sorry I need some sleep now. Thank your for your help with the GPS and with this S-pen so far. 👍😄 | 01:24 |
| T42 | <edp_17> Have a good night. | 01:24 |
| T42 | <elros34> look here: https://github.com/elros34/qtbase/blob/2b1cff97d24077e572ccb312c02d60b72f25ccd1/src/platformsupport/input/evdevtablet/qevdevtablet.cpp#L164 Events should be reported to compositor only if you have SYN_REPORT in evdev_trace so check it. Good night | 01:25 |
| T42 | <edp_17> @elros34: Thanks! There is SYN_REPORT but is always 0 (might be this is not a problem) | 08:07 |
| T42 | <edp_17> Do I need to write a program to test the pen or how else I can do it? | 08:08 |
| T42 | <edp_17> Might be a similar what you've done for the mouse cursor qxcompositor? | 08:26 |
| T42 | <W %lastname%> Hi, everybody. Does anyone know how to run anbox on sailfish os? | 08:29 |
| T42 | <elros34> @edp_17: 0 is ok. If i get it right, events are reported only if BUTTON_TOUCH != 0. Do get tht event when you press pen to screen? Test program could be useful too but according to qt doc, not handled QtabletEvent should be automatical converted to mouse events | 09:59 |
| attah | Interesting find (from the changelog) https://forum.sailfishos.org/t/release-notes-pallas-yllastunturi-3-4-0/2258/81 | 10:48 |
| attah | Is this a nod to community AOSP8 ports? | 10:48 |
| attah | It was fastboot that was flaky with USB3 right, any particular command, or all of them? | 14:20 |
| T42 | <adampigg> attah: all of them when using the pro1 .... i have to install a crappy old usb hub to get things working! | 14:38 |
| attah | @adampigg: i just managed to build fastboot from source without have any Android build env | 14:39 |
| T42 | <adampigg> attah: what are you working on ... printing not enough any more? | 14:40 |
| attah | and i thought it was working better (due to being newer), but it doesn't | 14:40 |
| attah | printing is so 1990 | 14:40 |
| attah | at least now i can strace it and see if a few more retries at certain places would help | 14:41 |
| attah | and i honestly don't know what i'm working on... just putting of cleaning my apartment | 14:41 |
| T42 | <adampigg> im just sitting down, having installed a new sink and removed the old ... all new pipework too | 14:43 |
| T42 | <adampigg> and my hands are covered in solvent glue :/ | 14:44 |
| attah | see, computers are much better :P | 14:44 |
| T42 | <elros34> After reboot neither gpsinfo nor puremaps nor messwerk can activate gps (flashing location icon). I have to use csd or toggle gps in settings. Looks like jolla-settings sets gpsFlightMode = false which is Powered property in /net/connman/technology/gps. WIthout Powered=true geoclue-hybris doesn't start positioning. Am I the only one w | 15:42 |
| T42 | ith this issue? | 15:42 |
| mal | does logcat still show the same errors | 15:43 |
| T42 | <elros34> mal: I guess you mixed it with edp_17 issue. This is different issue | 15:44 |
| piggz | mal: what chooses actdead.target ? | 15:50 |
| piggz | fixed it with | 16:21 |
| piggz | [root@VollaPhone nemo]# cat /var/lib/environment/actdead/bootparam.conf | 16:21 |
| piggz | ACTDEAD_PARAMETER_STRING='boot_reason=1' | 16:21 |
| mal | piggz: dsme | 16:53 |
| mal | piggz: and yes, that config file is the correct place to fix if it's working by default | 16:54 |
| piggz | mal: added check for boot_reason=1 as the defaults didnt work | 16:59 |
| piggz | working fine now .... charging is a little weird.... when you boot by charging, it loads up to the charging screen, but as soon as you unplug it powers off | 17:00 |
| mal | piggz: usually there should a small delay before shutting down | 17:01 |
| piggz | yeah, usually devices power down nicely, this just kills it! :D | 17:01 |
| mal | @elros34 sorry, so what do logs say? | 17:01 |
| mal | @elros34 is hw-settings.ini correctly set, is geoclue plugin installed and starting? | 17:02 |
| mal | @elros34 which android base and which geoclue plugin (hybris or binder)? | 17:03 |
| T42 | <elros34> mal: gps works fine. It's cm11 base, hybris plugin, latest version. The only problem is after reboot. If I start gps info I don't have flashing location icon. After boot gpsFlightMode is true and starting gpsinfo doesnt set gpsFlightMode to false. Toggling gps switch in jolla-settings helps because it set locationSettings.gpsFlightMod | 17:07 |
| T42 | e = false | 17:07 |
| T42 | <elros34> based on yesterday debugging of @edp_17 gps (different issue) he also had similar situation after reboot (IIRC cm14 base) | 17:09 |
| mal | @elros34 the toggle is the one setting the value of gpsFlightMode | 17:09 |
| T42 | <elros34> toggle in jolla-setting change both: locationSettings.locationEnabled and locationSettings.gpsFlightMode | 17:11 |
| mal | the question then is why is the value not stored properly, do you see any dconf issues in logs? | 17:13 |
| mal | did that work before? | 17:13 |
| T42 | <elros34> I am not sure whether it worked before, for sure it's same on 3.3.0.x and 3.4.0.22. I am not sure this is dconf. gpsFlightMode is read from connman via dbus (Powered property) | 17:16 |
| mal | hmm | 17:31 |
| mal | the state should be stored somewhere | 17:32 |
| T42 | <elros34> Don't know where. After toggling gps switch and using and closing gpsinfo, connmanctl technologies shows that GPS is powered | 17:35 |
| mal | @elros I think /home/nemo/.local/share/system/privileged/connman/settings | 18:58 |
| mal | hmm, it seems that is not changing | 19:00 |
| T42 | <elros34> this file shows only current status of wifi/gps. | 19:09 |
| T42 | <elros34> or maybe not. I mix up /home/.system/var/lib/connman from 3.3.0.x | 19:14 |
| mal | yes, there seems to both of those on device | 19:16 |
| attah | piggz: https://github.com/attah/my_fastboot | 20:35 |
| attah | not that it solves the problem... but at least it builds cleanly (: | 20:36 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!