| T42 | 武凯 %lastname% was added by: 武凯 %lastname% | 08:30 |
|---|---|---|
| deathmist | mal: you were working on AOSP ganges Pie bringup, right? well, the droid stuff (frameworks/native anyway) now needs to be updated and tested on 9.0.0-r60 so droidmedia can build again for Pie bases (hybris-16.0) in general: https://github.com/sailfishos/droidmedia/pull/70 | 08:42 |
| abranson | I think they did something similar in the early releases of android 7. no release discipline, that google :D | 08:45 |
| deathmist | it seems duplicating getLayerDebugInfo() in external/droidmedia/services/services_9_0_0.h below the original one and removing const from it works as well, can someone building AOSP (Sony) stuff on 9.0.0_r46 test if that breaks the build for them? | 09:32 |
| deathmist | https://github.com/JamiKettunen/droidmedia/commit/8c226af9e7199797e2c468eab0e3b7f2675d69c4 | 09:38 |
| deathmist | if that doesn't break anything it should be the optimal fix as nothing has to be rebased then Android HAL wise in anyone's tree :p | 09:43 |
| *** ghosalmartin is now known as mgrover | 10:05 | |
| mal | deathmist: that seemed to build without issues for ganges | 11:12 |
| deathmist | mal: neat, could you review https://github.com/sailfishos/droidmedia/pull/70? :) | 11:48 |
| PeperJohnny | hello! the hadk is for android phones? Or is it adaptable to linux devices as well? I'm asking because I have a pinetab and would like to bring sailfish to it | 16:12 |
| PeperJohnny | the current image has not the latest driver for the display and stays blank unfortunately | 16:12 |
| piggz | PeperJohnny: have you tried the kernel from the repo | 16:15 |
| PeperJohnny | the pine repo? | 16:15 |
| piggz | http://repo.merproject.org/obs/nemo:/devel:/hw:/pine:/dontbeevil/sailfishos_latest_i486/armv7hl/ | 16:16 |
| PeperJohnny | Do I get this one if I use the flash-it script? | 16:17 |
| piggz | hmmm | 16:17 |
| piggz | do you have a serial cable? | 16:17 |
| PeperJohnny | Because then I tried it, as my build was from 9.9 and it didn't work | 16:17 |
| PeperJohnny | nope. I could get one tomorrow earliest | 16:18 |
| piggz | i mean the uart lead that goes in the headphone jack of the tab | 16:18 |
| PeperJohnny | yeah. A colleague has one at work iirc. I could get that tomorrow | 16:19 |
| PeperJohnny | It's from an UPS but I guess it'll work? | 16:19 |
| PeperJohnny | Or do I need the pine one? | 16:19 |
| piggz | you could copy the kernel package to the sd card, boot, go in via the uart, and install the kernel | 16:19 |
| piggz | probably | 16:19 |
| PeperJohnny | Ok I'll try that tomorrow | 16:20 |
| PeperJohnny | I don't know if it's in the kernel tho. On the pine repos I saw that they had a special kernel with 5.7 with the drivers | 16:20 |
| PeperJohnny | arch ARM uses the 5.7 kernel too, even though 5.8 is the current one in the repos | 16:21 |
| piggz | i dont think we build the dts for the tab either | 16:21 |
| piggz | so thats probably another issue | 16:21 |
| PeperJohnny | https://gitlab.com/pine64-org/linux/-/tree/pine64-kernel-5.7.y-pinetabpanel | 16:22 |
| PeperJohnny | that's the one I'm talking about | 16:22 |
| piggz | you can build that manually and put it in the boot partition of the sdcard | 16:22 |
| piggz | just need the cross compiler setup | 16:22 |
| PeperJohnny | Or I could compile it on the tab itself :D | 16:23 |
| piggz | see https://wiki.merproject.org/wiki/Adaptations/PinePhone64 | 16:23 |
| PeperJohnny | which is probably a world of hurt in terms of time needed | 16:23 |
| PeperJohnny | From Installation on I guess? | 16:24 |
| piggz | yeah | 16:25 |
| PeperJohnny | Ok thanks a lot mate! | 16:25 |
| ghowa | @deathmist: I think I got a build error because of the change in getLayerDebugInfo: | 19:10 |
| ghowa | external/droidmedia/services/services_9_0_0.h:164:14: error: 'MiniSurfaceFlinger::getLayerDebugInfo' hides overloaded virtual function [-Werror,-Woverloaded-virtual] | 19:10 |
| deathmist | my patch should fix it, did for me locally anyway :p | 19:11 |
| ghowa | Is that the only change? Because I got two other errors as well: | 19:14 |
| ghowa | system/update_engine/update_status_utils.cc:19:10: fatal error: 'update_engine/dbus-constants.h' file not found | 19:14 |
| ghowa | and | 19:14 |
| ghowa | frameworks/av/services/camera/libcameraservice/CameraService.cpp:2050:48: error: unused parameter 'kind' [-Werror,-Wunused-parameter] | 19:14 |
| deathmist | you can ask kalube about that `update_engine/dbus-constants.h` one, but iirc he just copied it from some LOS repo. that CameraService.cpp error however is telling me that you're building audioflingerglue which was deprecated in favor of pulseaudio-modules-droid-hidl for all devices | 19:17 |
| ghowa | Pulseaudio is the sailfish side, right? | 19:23 |
| deathmist | yes, you don't build that with Android HAL | 19:25 |
| ghowa | So I just ignore this error? | 19:26 |
| deathmist | which one and what command did you run? | 19:26 |
| ghowa | The one in CameraService | 19:27 |
| ghowa | Command: make -j$(nproc --all) hybris-hal droidmedia | 19:28 |
| deathmist | yes ignore that one, you should not be building audioflingerglue anyway these days | 19:28 |
| deathmist | oh, hmm | 19:29 |
| mal | looks like I never remembered to push the fix for that error: unused parameter 'kind' [-Werror,-Wunused-parameter] | 19:29 |
| deathmist | ah, when I was doing local build testing I still had "-Wno-unused-parameter" in my frameworks/av/services/camera/libcameraservice/Android.mk which I've been using as a hack for a long time and thought it got cleared when I resynced and didn't build afglue today :p | 19:31 |
| mal | deathmist: there is a simpler and nicer way | 19:31 |
| deathmist | for fixing the cameraservice error/warning? sure, like I said it's merely a hack and it's been "working fine"™ | 19:32 |
| mal | deathmist: https://github.com/mer-hybris/hybris-patches/pull/9 | 19:40 |
| ghowa | mal, deathmist: thanks for the help, much appreciated! | 19:42 |
| mal | deathmist: which device are building? should you be using afglue? there is better way to do it | 19:44 |
| deathmist | I don't need afglue anymore: https://github.com/sailfishos-oneplus5/droid-config-cheeseburger/commit/58755cc772297d262c6d2d4724e386022265b035 | 19:44 |
| mal | ah, ok | 19:44 |
| mal | good | 19:45 |
| mal | that dummy af is the proper way to do it now | 19:45 |
| mal | deathmist: did you have issues with mobile data on that device? | 19:46 |
| deathmist | indeed, though adding & enabling dummy_netd fixed it: https://github.com/sailfishos-oneplus5/droid-config-cheeseburger/commit/3a353ed67c75f908889efda3c9308a6ed159aff9 | 19:46 |
| deathmist | this is strictly on 16.0, I didn't see this on 15.1 | 19:47 |
| mal | ok, good that you found that | 19:47 |
| mal | deathmist: why do you add that symlink? it's already in the package https://github.com/mer-hybris/dummy_netd/blob/master/rpm/dummy_netd.spec#L30 | 19:48 |
| mal | hmm, looks like I need to do some fixing to that dummy_netd for future sfos releases | 19:49 |
| deathmist | I don't remember exactly, I'll have to check as it could be useless actually then | 19:51 |
| deathmist | btw does https://github.com/sailfishos/droidmedia/pull/70 look ok now? | 19:52 |
| mal | deathmist: I did wonder if that should have been ifdef'd based on version but maybe not needed as that does work now | 19:54 |
| T42 | <ItsMeShouko> I am getting this error related to droidmedia when trying to build android hal, it was not coming before... https://paste.linux.pizza/?da392f604834e446#4PBL8BLdNNdgwdbtjkTWhzY6F1Z5RyrKGvBuyjm6F5f7 | 19:55 |
| deathmist | @ItsMeShouko look at the PR I literally just sent lol | 19:56 |
| T42 | <ItsMeShouko> oof, nice | 19:57 |
| mal | deathmist: do you have any issues on that device which you haven't solved yet? | 19:59 |
| deathmist | there's some weird camera behavior (i.e. corrupted video) and flashlight activates in a strange way (it's a bit delayed and flashes rapidly a few tunes when disabling every time) | 20:00 |
| deathmist | for Bluetooth audio I think I need the BT_SCO dbus-send thing which seems to be an MSM8998 specific quirk which all devices need, but that I haven't tested just yet | 20:00 |
| deathmist | dbus-send --print-reply --address='unix:path=/run/user/100000/pulse/dbus-socket' --dest=org.PulseAudio1 /org/sailfishos/audiosystempassthrough org.SailfishOS.AudioSystemPassthrough.set_parameters string:"BT_SCO=on" | 20:01 |
| deathmist | ^ that's the one I'm talking about | 20:01 |
| mal | ah, I think there was some other device with similar hack needed | 20:02 |
| mal | deathmist: how does the recorded video look like? | 20:02 |
| deathmist | I' | 20:03 |
| deathmist | ll send a preview in a moment | 20:03 |
| deathmist | also about that BT_SCO thing, it's for call audio routing specifically; media playback for example works fine already | 20:04 |
| T42 | <ItsMeShouko> is there any news of hybris-17? | 20:12 |
| deathmist | I just tested the BT_SCO thing and it works, as in the call audio is properly routed to/from the Bluetooth headset instead of the phone pretending it isn't connected at all during calls :) now just need to make some unit to run that on boot | 20:49 |
| deathmist | mal: I'll send a video preview tomorrow, it's late now and I can get a better comparison in daylight anyway | 20:50 |
| deathmist | basically the first 2 seconds seem to be "corrupted" every time and the video has "compression artifacts" all over the place when moving even slightly, on LOS 16.0 base ROM this doesn't happen | 20:51 |
| deathmist | https://github.com/sailfish-on-fxtecpro1/droid-config-t5/commit/9abbe7f58f6f869a3ee3736bf91c38b56f28accb seems people did the work already for MSM8998 devices :D | 20:53 |
| mal | deathmist: yes, that was the device I was talking about | 20:55 |
| T42 | <ItsMeShouko> I tried enabling home encryption in my build but when sfos boots it says could not encrypt user data and then shows black screen and stays there... Is there anything that I can try? | 22:35 |
| T42 | <NotKit> @ItsMeShouko is your build LVM-based? | 22:37 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!