Monday, 2020-11-30

T42<birdzhang> Chengcheng: Hi, most porters are from Europe, they are sleeping now :)02:11
rinigusmal: morning! looks like webgl2 started to work in sfos 3.4. tested in stock and angelfish. while stock browser was updated between 3.3->3.4, angelfish and runtimes are the same06:42
rinigusmaybe it is an improvement brought by gles306:42
rinigusI wonder whether we have a device updated to 3.4 but without updating libhybris (or not supporting gles3)...06:43
malrinigus: hmm, not sure what webgl2 needs, is there some test page for that?08:40
rinigusmal: I used
rinigusthere is also that I will try tonight08:57
T42<neochapay> rinigus ping in nemo ^_^09:13
malrinigus: what was the libhybris version you used before?09:14
rinigus@neochapay: replied in nemo already. I wonder if the bridge is broken :(09:35
T42<neochapay> oh......09:35
rinigusmal: according to OBS, SFOS 3.3 was using; 3.4 is using
malhmm, that version should have been fine09:37
rinigus@neochapay: just in case - I had no progress on nemo for ~1.5 months. taking a, hopefully, short leave to work on other projects09:38
T42<neochapay> not good09:38
malmaybe I need to test on 3.3 build for fp209:38
rinigusmal: maybe. assuming that with 3.4 you get webgl2 test working09:39
malrinigus: works on 3.4 fp2 at least09:40
rinigusmal: that's consistent with sony. if you have time or maybe someone from fp2 users on 3.3 could just test it09:41
malrinigus: I can easily flash my second fp2 device with 3.3 later09:42
rinigusmal: thanks, looking forward09:42
malrinigus: from wikipedia "WebGL 2.0 is based on OpenGL ES 3.0"09:44
rinigusmal: so, do we have then the test for gles3 on working systems? assuming that gl es 3 was added after hybris .3409:46
malrinigus: opengl es 3 support was added in libhybris 0.0.3509:49
malrinigus: it seems I had .36 in 3.3.0 release for fp209:50
rinigusmal: so, currently we have test done on sony tama (3.3 on xz2 vs 3.4 on xz2c) with the hybris before and after gles3 support. it is probably sufficient for understanding what has happened, but maybe some other ports could check it.09:53
T42Ralph Black %lastname% was added by: Ralph Black %lastname%11:41
T42<Andy %lastname%> Bit of mixed progress one the OnePlus 3T - not been an update for a while with an SFOS image built after delaya73 . There was a more recent build back in April I think here
T42<Andy %lastname%> Anyway this had a working camera but busted bluetooth.17:29
T42<Andy %lastname%> I have upgraded via the terminal to since and of course this drives a wreck through the phone a little (losing the screen viewfinder for taking photos - which still take BTW.  This weekend I had a little time and also tried to upgrade but pointing to upgraded community files on Mer.17:30
T42<Andy %lastname%>
T42<Andy %lastname%> Conclusion - upgraded again to  but no worse than with and if anything a little faster on
T42<Andy %lastname%> Question...How hard would it be to restart builds for this? Using the pipelines from before?17:33
T42<elros34> @NotKit do you have any pointers how to rotate display in qt5-qpa-hwcomposer-plugin for external display? I set HWC_TRANSFORM_ROT_90 for layer 0 and 1 but either there is no effect or I have garbage on display. The only way external display work is when I set it to be same resolution as primary display.17:41
malrinigus: ping17:42
rinigusmal: pong18:45
malrinigus: I noticed that the accelerator data should be filtered already and
rinigusso, I guess suspicion is that there is no need to filter further at compass level?18:52
rinigusmal ^18:52
malnot sure18:53
malrinigus: at least filtering the accelerometer input like you suggested doesn't make sense, the current way of filtering the final value might make sense18:54
rinigusmal: I would argue that angle and fBf* filtering are equivalent. not numerically, but by idea behind it18:56
malrinigus: we already have filterin for both inputs, mag and accel18:57
malrinigus: magDataAvailable does it for mag and for accel it's done in the avgaccfilter18:58
rinigusmal: yes, I can see that. but maybe then compass should be free from filtering on it's side. I wonder how unstable it will be18:58
malI can test how it behaves18:58
malI think it updates a bit too often18:58
rinigusmal (see that - > from your references and code above the filtering)18:58
rinigusmal: but update frequency should be not dependent on that...18:58
malupdate frequence depends on accel data input rate18:59
rinigusmal: yes, but will stay the same regardless to the filtering19:04
malrinigus: we could add another downsamplefilter after the compassfilter19:06
rinigusmal: yes, that would probably work. I wonder whether Qt also has some filtering/downsampling on top of it all19:09
rinigusI have a problem with OTA to 3.4 from 3.3. during update, mce is closed, started again, but it fails to start. can't start mce later either. as a result, you end up with the device having a screen off and no way to start it19:20
rinigusI am logged in into device currently via ssh19:21
rinigusin logcat, I am getting messages "Opening '/dev/binder' failed: Operation not permitted"19:29
maldid the update finish or still happening?19:30
rinigusmal: it finished. I can poke around19:32
rinigusI presume that if I reboot, all be fine. have seen it once on smaller update and that's the way it went19:33
rinigus... that time. so, maybe I can learn something before reboot19:33
riniguswhat's strange: crw-rw-rw-    1 root     root       10,  53 Nov 30 20:39 /dev/binder19:36
rinigusand at the same time: 11-30 19:06:24.068  2306  2306 W ProcessState: Opening '/dev/binder' failed: Operation not permitted19:36
rinigusselinux strikes again?19:37
rinigusno idea even what to check, unfortunately19:39
rinigusMaybe it is mismatch in some package versions. Will follow all versions in sfos changelog and try again19:51
rinigusyep, reboot fixed it. not the best user experience for OTA, though19:56
malrinigus: I need to test on fp2 to see how it behaves on that20:22
rinigusmal: I am going through packages meanwhile to be sure that nothing got updated too far20:23
malrinigus: but that error looks quite odd20:23
rinigusmal: it does20:24
wdehoogis it correct that for /home/nemo does not exist anymore? and now it is /home/defaultuser?20:28
malwdehoog: in new images yes, in old images updated to new version user is still nemo20:28
wdehoog@mail aha. someone has problems with email and the log shows /home/nemo but I installed an image and could not find it20:29
wdehoogthis user did OTA and now email does not work anymore20:30
wdehoogI (reinstalled image) can add a new account and it works. he cannot, old accounts fail and new accounts fail.20:31
malhmm, logs would be useful20:32
wdehoogyup but that is a bit problematic. see here
wdehoog@mal more logs at looks like messageserver5 crashes (what is it?)21:07
wdehoogemail works thanks to this:
rinigusmal: same issue with crashed binder on a second try. this time used exactly the same versions as in the release changelog. let me know how it goes with fp221:40
riniguspiggz cheated/went around ota, as far as I understood. wonder what's experience for others with 3.3->3.4 ota. although, will catch up tomorrow, late already21:41
piggzwhat did i cheat?21:42
malrinigus: did you get full journal log from the update?21:50
T42<Andy %lastname%> @wdehoog [I (reinstalled image) can add a new account an …], I had to reinstall / reflash with the same  issue.21:52
rinigus:) piggz not much, I have odd issue with update.21:53
piggzmal: what is this ?21:53
rinigusmal: have a log from previous one, will upload tomorrow. Journal and logcat21:54
wdehoogAndy: you mean email problems? the link to the forum shows how it can be solved without reflashing21:54
wdehoogAndy: does vibrator work for you? If not I have updated droid-hal-bacon in the testing repository so if you update it might work.21:55

Generated by 2.17.1 by Marius Gedminas - find it at!