*** OhYash1 is now known as ohyash | 08:09 | |
T42 | <DylanVanAssche> Anyone who knows why sensorfw won't load his plugin? | 12:07 |
---|---|---|
T42 | <DylanVanAssche> ```Oct 13 14:00:18 Sailfish sensorfwd[2479]: Loader loading plugin: "accelerometerchain" as: "accelerometerchain" from: "/usr/lib/sensord-qt5/libaccelerometerchain-qt5.so" | 12:07 |
T42 | <DylanVanAssche> Oct 13 14:00:18 Sailfish sensorfwd[2479]: Plugin not available: "accelerometerchain" | 12:07 |
T42 | <DylanVanAssche> Oct 13 14:00:18 Sailfish sensorfwd[2479]: plugin marked invalid: "accelerometersensor"``` | 12:07 |
T42 | <DylanVanAssche> The .so file exists though... | 12:07 |
T42 | <DylanVanAssche> I checked the hw-settings file from CSD, all sensors except for the GSensor (=accel) are disabled, if I check ssu-sys-info, it lists more sensors as features than enabled :( | 12:09 |
vknecht | @DylanVanAssche: main config file would rather be droid-configs/droid-configs-device/sparse/etc/sensorfw/primaryuse.conf and file(s) in sparse/etc/sensord.conf.d/ I guess | 12:19 |
T42 | <DylanVanAssche> vknecht: It uses the hardware features to detect the sensors, these are all the files there: https://bin.disroot.org/?42ae0dedab37f097#53r5dgte1AbkDFY2ksiXiQOTjCHYpu/Rty3W3XRGSME= | 12:24 |
T42 | <elros34> it should be iiosensorsadaptor not iiosensoradaptor | 12:27 |
T42 | <DylanVanAssche> @elros34 Oh seriously, missed that one again :O | 12:31 |
T42 | <DylanVanAssche> For some reason my accel sensor is saturating between min and max range. | 12:32 |
T42 | <DylanVanAssche> `Acc value discarded due to over/underflow` | 12:32 |
T42 | <elros34> maybe it should be iiosensoradaptor then, I just used commented out name | 12:33 |
T42 | <DylanVanAssche> So there must be something wrong with the calculation (raw, scale, ...) | 12:33 |
T42 | <DylanVanAssche> the adaptor supports more than 1 sensor (all IIO sensors) so 'sensors' seems fine for me | 12:34 |
T42 | <elros34> Didn't you changed min/max range? | 12:35 |
T42 | <elros34> this also deosn't looks correct for me https://github.com/sailfish-on-dontbeevil/sensorfw/commit/a06f64296358d53859655ff78f22bdc99e4ea6f8 as it worked for others | 12:37 |
T42 | <DylanVanAssche> That was an experiment, might reverse it later. If I calculate manually the (result + offset) * scale I get the same reading as Messwerk on my Xperia X. So without the *1000 and REV_GRAVITY, the value should be correct. | 12:39 |
T42 | <DylanVanAssche> I thought that that last part was the cause of the saturation | 12:39 |
T42 | <elros34> as a comparison: dmard06: 2g accelerometer: -33:+33 raw value, scale 0.031250 so it's -1:+1 both when reading from sysfs and in messwerk (m/s^2). Not sure if these values are correct but rotation works correctly in glacier | 12:52 |
T42 | <DylanVanAssche> I get intermediate values when reading from sysfs. Seems that the scale is interpreted wrong. When setting up the sensor, I have 0.306457, when reading the sensor, the scale is 1. I'm reverting my debug commits now and will check if it's not related to one of those commits | 12:55 |
T42 | <DylanVanAssche> Well... reverting the commits fixed it :O | 13:00 |
T42 | <DylanVanAssche> Seems that now the REV_GRAV is needed. It didn't worked before with it, weird! | 13:00 |
T42 | <DylanVanAssche> Now fixing rotation... | 13:01 |
T42 | <elros34> Maybe the calculation in sensorfw are really wrong because messwerk shows values +-0.1g and I think it should +-1g | 13:13 |
T42 | <DylanVanAssche> Z value in messwerk should be -1g | 13:14 |
T42 | <DylanVanAssche> when you put your dev on a flat surface | 13:14 |
T42 | <DylanVanAssche> *1g = 10 m/s^2 | 13:15 |
T42 | <elros34> yes, z has 1g when reading from sysfs but in messwerk it's 0.1g | 13:18 |
T42 | <DylanVanAssche> Well not on my device... | 13:18 |
T42 | <DylanVanAssche> You need to scale up the value then? Sailfish on Latte has something similar: https://github.com/sailfish-on-latte/droid-config-latte/commit/6c04db5e30c791002905f284412aeb56d7917d3c (scales down with a factor of 1000) | 13:19 |
T42 | <DylanVanAssche> With only an accel sensor, you can rotate the screen in landscape and vice versa I think (according to what I understand from the default sensorfw config: `orientationsensor=Feature_GyroSensor|Feature_AccelerationSensor`). | 13:20 |
T42 | <DylanVanAssche> *? | 13:21 |
T42 | <elros34> sure I have no issues with rotation | 13:22 |
T42 | <DylanVanAssche> Well rotation is Messwerk doesn't show anything (no graph) | 13:23 |
T42 | <DylanVanAssche> Aha, it works anyway :) just no data in Messwerk. I guess it only shows data when you have a gyroscope (where would it otherwise get angles from) | 13:25 |
T42 | <adampigg> Yes,.i think thats how it works in latte too | 13:30 |
T42 | <DylanVanAssche> Well we have almost a fully working tablet then :) just sound and a flakky charging driver | 13:31 |
T42 | <elros34> have you done anything special to make lipstick works with eglfs? For me it always crash in closed part (/usr/bin/lipstick), glacier and language selection works. | 13:35 |
T42 | <DylanVanAssche> @adampigg You know more about eglfs and Lipstick crashing? | 13:36 |
T42 | <adampigg> What sfos version? | 13:38 |
T42 | <adampigg> There was a fix in 3.1 | 13:38 |
T42 | <adampigg> Wasnt a closed part tho | 13:39 |
T42 | <elros34> it fails on 3.1. I wonder whether it could be related to native landscape display or very small resolution | 13:41 |
T42 | <adampigg> It could have been 3.1.1 that had the fix | 13:42 |
vknecht | there's no unique path to access sdcard, is there ? /media/sdcard/ always has sdcard-specific UID as a subdir it seems | 13:55 |
vknecht | can we at least rely on /media/sdcard always being sub-root of (all ?) sdcard(s) ? | 14:05 |
T42 | Bob_Da_Builder was added by: Bob_Da_Builder | 15:52 |
T42 | <DylanVanAssche> We have 3 subprojects on Mer OBS for device specific packages like droid-config-pinephone. | 16:15 |
T42 | <DylanVanAssche> 2 of them (pinetab + devkit) are building fine. The 3rd one: droid-config-pinephone is missing `android-headers`. I don't understand why this package won't build since it's almost the same as the other 2. | 16:15 |
T42 | <DylanVanAssche> https://build.merproject.org/package/show/nemo:devel:hw:pine:dontbeevil:pinephone/droid-config-pinephone | 16:15 |
T42 | Udhayugi was added by: Udhayugi | 17:03 |
lbt | PSA: git.sailfishos.org (and git.merproject.org) should be back to normal. Let me know of any issues. | 17:12 |
T42 | <DylanVanAssche> lbt: seems faster :) | 17:15 |
lbt | good! | 17:15 |
Thaodan | vknecht: why not use run media? | 17:40 |
vknecht | what do you mean ? btw, just PRed my wip proposal :-) https://github.com/piggz/harbour-advanced-camera/pull/78 | 17:42 |
vknecht | detects if there's only one dir in /media/sdcard, and uses that if true | 17:43 |
vknecht | Thaodan, ah, /run/media ? well, there's still the sdcard UID to account for eg. /run/media/nemo/0403-0201/ | 17:51 |
vknecht | wondering, are there devices with multiple sdcards slots ? | 17:52 |
Thaodan | yes | 17:53 |
vknecht | hum, there will be some sport then, I guess :-) | 17:53 |
Thaodan | the dir is for any connected media such as otg | 17:54 |
T42 | <elros34> more likely you can have multiple partitions than multiple sdcards | 17:55 |
vknecht | ah yes, too bad I just found my usb otg key in the washing machine, won't be able to test | 17:55 |
vknecht | uh, yes | 17:55 |
vknecht | maybe better/simpler to only offer picking a specific dir rather than trying being smart ? | 17:57 |
piggz | mal: around for some camera reviews | 18:27 |
piggz | ? | 18:27 |
mal | piggz: I'll try, feeling a bit tired though | 18:58 |
mal | piggz: should I remove the packages that are in native-common from dontbeevil OBS projects, no need to have duplicates | 18:59 |
T42 | <DylanVanAssche> mal: If they are upstreamed, I don't see any issue there. | 19:02 |
T42 | <DylanVanAssche> Are these packages still build from the Github sailfish-on-dontbeevil project or did you fork them somewhere? | 19:02 |
mal | @DylanVanAssche some of those in native-common are from sailfish-on-dontbeevil project | 19:03 |
mal | @DylanVanAssche btw, the iio sensor changes were merged already to upstream, not the wip proximity changes of course | 19:04 |
T42 | <DylanVanAssche> mal: Okay, just to make sure that we don't clean up our Github project and some packages are losing their source repo :P | 19:04 |
mal | many of the other changes have not been upstreamed yet | 19:05 |
T42 | <DylanVanAssche> mal: I saw it :) I got the PineTab accel sensor working too with these changes. I will revise my proximity sensor changes soon and see if I can get it working :) | 19:05 |
T42 | <DylanVanAssche> other changes? | 19:05 |
mal | I mean those other packages in there | 19:06 |
T42 | <DylanVanAssche> Oh yeah, sure :) | 19:06 |
mal | for example the v4l stuff | 19:06 |
mal | @DylanVanAssche does pinephone already use the new pulseaudio-module-keepalive instead of the old hacky pulseaudio-modules-droid-keepalive | 19:07 |
T42 | <DylanVanAssche> Haven't got it working in Qt :( | 19:07 |
mal | that needs some pulse config changes to use pulseaudio-module-keepalive | 19:08 |
mal | those are in droid-config submodule | 19:08 |
T42 | <DylanVanAssche> mal: I didn't checked that out yet. We use the default linux pulse audio config, not the one from Jolla ;P | 19:08 |
mal | we really should do a native port submodule | 19:08 |
T42 | <DylanVanAssche> Probably :) | 19:08 |
mal | @DylanVanAssche how does that handle the keepalive stuff? | 19:09 |
T42 | <DylanVanAssche> mal: I have no idea :O | 19:09 |
mal | https://github.com/mer-hybris/droid-hal-configs/commit/0f656df7d653cb5b45ea575f61c58f1121eeae20#diff-22c2dfe7632bad14b13275be41255a4b | 19:09 |
T42 | <DylanVanAssche> I haven't tested it properly yet, just some tunes from the Settings app. Audio is broken on the PineTab though | 19:09 |
mal | that is how the config loads keepalive | 19:10 |
mal | ok | 19:10 |
T42 | <DylanVanAssche> I can test it on the PinePhone dev kit. I haven't got a real PinePhone yet :( | 19:10 |
mal | sure, I of course meant the dev kit | 19:10 |
T42 | <DylanVanAssche> Yeah I thought so :) | 19:11 |
T42 | <DylanVanAssche> For the sensors stuff is a real device a bit easier though :P | 19:11 |
*** OhYash1 is now known as ohyash | 19:26 | |
piggz | back! | 19:38 |
piggz | mal: is our new obs structure ok with tou | 19:39 |
piggz | you | 19:39 |
mal | piggz: a bit odd maybe, usually different device are parallel not under some device | 19:54 |
mal | piggz: I can add new devices if needed | 19:55 |
piggz | we have a lot of shared packages ... only need config and version in device specific | 19:56 |
mal | piggz: aren't most of those shared ones common to all native ports? | 19:57 |
piggz | hmmm, quite possibly | 19:57 |
mal | piggz: but you can also have something like this if needed https://build.merproject.org/project/show/nemo:devel:hw:semc:common | 19:58 |
mal | piggz: I'll cleanup that dontbeevil project now to remove those that are in native-common | 19:58 |
mal | @adampigg a small issue in the rotation PR (well not directly that code), if I touch outside the menus when menu is open then the focus is set to that, I think the focus point shouldn't change when menu is being closed | 21:26 |
mal | hmm, seems to be a bit random when that happens | 21:27 |
mal | @adampigg also the not sure if the face-up and face-down orientations should be handled at all, in my opinion in those case we should keep the previous orientation i.e. not handle at all | 21:45 |
T42 | <Umeaboy> Does anyone know what Jolla has said about making a x86 version of Sailfish? There are laptops with touchscreens. | 22:53 |
T42 | <NotKit> there unofficial x86 ports (besides Jolla Tablet), but I doubt it makes sense for them to maintain generic x86 version for those kind of laptops | 22:58 |
T42 | <NotKit> (for example Xiaomi Mi Pad 2) | 22:58 |
T42 | <Umeaboy> Okey. | 23:00 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!