Thursday, 2019-06-27

liuxjusa, I've got complete pulseaudio log with "-vvvv" added to pulseaudio.service. could you help check it if you have time? log pasted to:
Mister_Magistermccreary: nevermind though but thanks04:50
Mister_Magisterlooks like mal just went to sleep :P04:50
T42<adampigg> R0kk3rz, mal, abranson, any thoughts.....flashed hybris boot, phone boots into recovery with telnet on 23....05:55
T42<adampigg> I format and extract the rootfs onto userdata05:56
T42<adampigg> No longer boots, doesnt even show on dmesg05:56
T42<adampigg> All i can think of is to flash stock boot, load android and try amd get logs05:57
r0kk3rzrecovery? as in twrp?05:57
T42<adampigg> There is no recovery yet05:58
T42<adampigg> To flash the rootfs, we are doing it thru telnet manually05:59
r0kk3rzwhat device is this?06:00
T42<adampigg> Fxtecpro106:00
r0kk3rzah ok06:00
r0kk3rzso you get telnet?06:01
T42<adampigg> Not after extracting the rootfs06:01
T42<adampigg> I do before06:01
r0kk3rzso the switchroot phase06:02
T42<adampigg> Yes06:02
T42<adampigg> I say 'I' as it works for mister_magister!06:02
T42<adampigg> Currently reading the initscript06:03
r0kk3rzso chen has been delivering the goods06:03
T42<adampigg> Yes, hes been spot on06:04
T42<eugenio_g7> @adampigg using LVM or stpwawaying over Android?06:04
T42<adampigg> Regular port atm06:05
T42<eugenio_g7> Have you tried touching init_enter_debug2?06:05
T42<adampigg> Not yet,  it was late last night, and to do that i will have to boot android anyway06:06
T42<eugenio_g7> Ok06:06
r0kk3rzso in the telnet you should be able to mount the rootfs and make sure it mounts ok06:08
T42<adampigg> After setting init_enter_debug2?06:10
r0kk3rzjust mount it somewhere06:11
T42<adampigg> I dont get telnet atm06:12
T42<adampigg> Oh, you mean start again06:12
T42<adampigg> And, after extracting,  mount then06:12
Mister_Magister@adampigg congrats u got my nick :P06:13
T42<adampigg> Blah blah06:13
r0kk3rzis maybe where its explosing06:14
T42<adampigg> Yup, been reading it06:14
T42<adampigg> Ta06:14
Mister_Magisterr0kk3rz: doubt as he had to mount data to extract rootfs06:17
r0kk3rzwell i dont know what weird shit you guys are doing :P06:18
T42<adampigg> Data did look fine, was freshly formatted with a lost+found....06:18
T42<adampigg> Around line 390 where it switxhes maybe06:18
r0kk3rzdoes it has a leddy?06:20
Mister_Magisterrgb one06:20
Mister_Magisterand it's working ootb06:20
Mister_Magister@adampigg after extracting try starting it manually06:21
Mister_Magisterdo what the script does06:21
T42<adampigg> Yes, thats what i was thinking06:21
r0kk3rzso add some leddy triggers to the init script06:21
Mister_Magisterr0kk3rz: we can also use keyboard backlight *wink wink*06:22
Mister_Magisteroh i need to ask spiiroin about keyboard backlight06:23
T42<adampigg> All good debugging for later06:23
Mister_Magisteri'm not doing anything at work since 3 days so no debugging here either06:24
T42<adampigg> And how to enabke osk when keyboard away!06:24
Mister_Magisteryeah that too06:24
Mister_Magisterand how to turn screen on with keyboard06:24
r0kk3rzi thought TheKit had it booting already?06:24
Mister_Magisterr0kk3rz: different device06:24
Mister_Magisterlike, older model06:24
r0kk3rzoh ok06:24
r0kk3rzprototype rather than pre-production, cool06:25
Mister_Magisteryeah yeah something like that06:25
T42<iHusky> hi someone develop AOSP android because I had problem07:23
spiiroinMister_Magister r0kk3rz: support for keyboard backlight in sysfs was added by mal some time ago, ref:  -> pull req -> sample config08:13
merbotMer bug 1349 in mce "Add support for a simple user-defined keyboard backlight" [Task,Resolved: fixed]08:13
Mister_Magisterthank you08:14
deathmistI'm attempting to fix sensors (at least orientation, als & proximity) on 15.1, sensorfwd status journalctl dmesg logcat
T42<birdzhang> does sensors service running ?09:38
Mister_Magisterthat was my thought09:38
T42<birdzhang> and if running, what's test_sensors output09:40
deathmistpgrep sensor returns nothing, so assumably not09:40
Mister_Magister then better fix it :P09:42
deathmistsorry, I'm an idiot, I ran that on my host, sensor and sensors both return processes09:42
deathmisttest_sensors segfaults09:42
T42<birdzhang> @deathmist good, then use strace -f -o sensors.log test_sensors09:43
T42<birdzhang> and goto end of sensors.log09:43
deathmistit's trying to access a bunch of sensor libraries in /vendor/lib/hw and other directories, which don't seem to be there09:44
T42<birdzhang> i assume you need this
T42<birdzhang> change to yours09:45
maldeathmist: did you add the hw-settings.ini file ?09:48
Mister_Magistermal: o/ what's up!09:53
Mister_Magisteru went to sleep yesterday?09:54
r0kk3rzMister_Magister: i watched steins;gate the other week, was weird09:56
Mister_Magisterr0kk3rz: lol weird? not good?09:57
r0kk3rzi binged the whole thing in a weekend, so yeah it was good09:57
Mister_Magisterr0kk3rz: don't talk about time machine cause CERN will get u :P10:04
Mister_Magisteralso don't put bananas in microwave10:04
deathmistbirdzhang: I added symlinks and now test_sensors returns stuff \o/ mal: I don't have that file, saw it in faq so I'll work on that next10:04
T42<birdzhang> @deathmist :D10:05
Mister_Magister@birdzhang that actually might be useful to me10:06
Mister_Magisterit's for 15.1 right10:06
birdzhangyeah, 15.110:06
Mister_Magisterwell that didn't help me :P10:16
Mister_Magisternot again! /usr is not even in paths it's checking10:20
birdzhangMister_Magister: your $HYBRIS_LD_LIBRARY_PATH maybe empty, and you use latest dhc upstream10:27
Mister_Magisteri don't think its latest10:28
Mister_Magisteri did git reset --hard10:28
birdzhangtake a look at /usr/bin/droid/droid-hal-startup.sh10:29
Mister_Magisterye it sets path to empty what about that10:29
Mister_Magister thats my dhc10:30
birdzhangwell, it's okay when HYBRIS_LD_LIBRARY_PATH is empty, why /usr/libexe/droid not in paths10:34
Mister_Magisterwell kay but i didn't emptied it10:36
deathmistmal: after adding a hw-settings.ini as well orientation works! :) sensorfwd status returns no more missing plugins either10:56
deathmistbrightness works properly too now, it was too high still when going to lowest before10:57
Mister_Magistermal: u will have time after 16?11:00
malnot sure when I will stop working, probably much later than that11:03
Mister_Magistermal: just… you said yesterday that you will give me fix and i'm wondering when i can ask for it :P11:04
Mister_Magistermal: will you just write fix on pm to me when you have time for that?11:06
deathmiston my device /sdcard is a broken symbolic link to /storage/self/primary, could I just add a symlink to /data/media instead (proper internal storage path) to sparse files and expect it to work?11:26
deathmistso basically do sparse files also overwrite existing files on the rootfs? :p11:27
Mister_Magisterdeathmist: you might get conflic12:03
deathmistthat's what I'm thinking, will test in a bit but first gonna fix GPS too (same issue, no .so files in /usr/libexec/droid-hybris/system/lib/hw/)12:06
Mister_Magisterwhy do you need /sdcard anyway12:07
deathmistI don't, just a quick access thing on the cli :p12:07
Mister_Magistersailfish doesn't use it12:07
Mister_MagisterOKAY guys i have no sms/email notification sound confirmed by 2 other people. Any slightest idea how to debug/solve it?12:08
maljournalctl at least12:13
Mister_Magistermal: you didn't answer me yet ;-;12:14
malwell the answer is obvious12:14
deathmistso I don't think the OnePlus 5 has regular GPS at all since there are no GPS related libs in either /system or /vendor, just found gnss12:30
T42<birdzhang> @deathmist works with GPS too, link /vendor/lib/hw/android.hardware.gnss@1.0-impl-qti.so12:31
maldeathmist: gnss is gps12:32
malor to be accurate gnss includes gps, glonass and others12:33
Mister_Magister udev is drunk. it wants to chown something in loop but fails lol12:38
Mister_Magisterwhy udev this path doesn't exists12:39
deathmistmal: oh ok, well even with that file symlinked it still doesn't find a GPS interface, geoclue pkgs ptrace
T42<birdzhang> @deathmist test_gps is not working, but GPS apps show my location12:50
T42<birdzhang> this app is harbour-gpsinfo12:52
deathmistok, but also came across "/etc/gps.conf: broken symbolic link to /system/etc/gps.conf"12:53
maldeathmist: what symlink are you talking about?12:54
deathmistmal: /etc/gps.conf, I found the file in /vendor/etc/gps.conf instead12:56
deathmistwill try symlinking to etc12:56
deathmiststill nothing, will try an app13:01
malmesswerk or gpsinfo are good for that13:02
liuxthe "jolla-ambient-z$ICON_RES" in hadk cannot be downloaded anymore. Is there any repo to fetch the package for ratio settings?13:03
Mister_Magisterngfd segfaults?! what?!13:07
malliux: what are you trying to do?13:08
malthose packages should be in standard repos13:08
liuxmal, "to change pixel ratio on a running device" in hadk13:09
Mister_Magistermal: now i know why there is no notifications. ngfd segfaults13:12
malMister_Magister: which ngfd-plugin* version do you have?13:15
Mister_Magisteri have one, no multiple questions13:17
Mister_Magisterwhat does droid-vibrator do there13:17
Mister_Magisteri have installed droid-vibrator not native-vibrator for some reason13:19
Mister_Magisterlooks like i havent changed to native yet on this device. weird i though i have13:20
deathmistjust tried gpsinfo outside as well, device only doesn't seem to get a fix but the other options work fine13:23
malso you see satellites in the list?13:24
Mister_Magisterhah notification is there13:25
deathmistit just said 0/22 on satellites in use / view when I was outside too for a while13:26
maldeathmist: ok, then it just probably needs some time to get the data from the satellites before it can report a lock13:27
malso gps is fine for your13:27
Mister_Magisterthen gps is working i think, for making sure you should wait for just 1 satelite catch up13:27
malthat already reported 22 visible satellites which seems reasonable13:28
Mister_Magisteryeah but sometimes phone won't catch lock on satelites13:29
Mister_Magisteri had that problem once13:29
Mister_Magisteri was out in the city i have never been13:29
Mister_Magisterthat was fun to use map without actual gps13:29
deathmisthooray it just got a fix on device only \o/ 8/35 now near the window13:31
Mister_Magistermy moto g2 became fully 100% working again13:31
liuxMister_Magister, cong13:32
deathmistthat's my goal with this 1+5 port too, planning to use this as a daily driver :)  how is fingerprint sensor support in SFOS? I can't seem to find a setting in Settings > Device lock, just a security code13:34
Mister_Magisterdeathmist: boi 4 years ago i was using this same moto g2 as daily driver with like 50% of stuff not working xd13:35
Mister_Magisterdeathmist: not even jolla can manage their own package so i say it's bad, really bad13:35
malMister_Magister: so the ngfd issue was easy to fix it seems13:48
Mister_Magisteryeah super easy13:49
Mister_Magistergotta change to native today13:49
malalso check all of your other devices13:50
liuxmal, I've modified ratio to 1.4 and icon to 1.5. after rebooted, the app icons were crowed to the top center of the screen and keep small size.13:54
deathmistmal: aren't BT_HCISMD and BT_HCIVHCI needed for bluetooth on 15.1? I can't seem to get HCISMD to stick in .config, probably another missing dependency option right?14:32
Mister_Magisterdeathmist: u need to patch kernel for HCISMD but i'm not sure if it's needed in 15.214:35
deathmistwell I can't enable BT_HCIUART as it causes
maldeathmist: only latter i.e. BT_HCIVHCI14:41
deathmistmal: alright, since the option didn't stick anyways I'll try with the rootfs I'll have ready soon :)14:42
malBT_HCISMD is only relavant for hybris-14.1 or older14:53
deathmiststill no BT :/  defconfig .config hw-adaptation.yaml how would I make sure bluebinder is active?15:02
deathmistit definitely is installed on the device right now15:03
malis bluebinder running?15:05
deathmistit's failing
liuxmal,  need your advices on app icons issue:
liuxdconf write /desktop/sailfish/silica/theme_icon_subdir "'z1.5'"15:12
deathmistHOLY CRAP I just got BT working probably fully by just enabling BT_BNEP and flashing the new boot image!15:59
deathmistok BT audio doesn't work (just checked on a pulley menu)16:00
deathmistnvm, media audio works from YT for example!16:03
maldeathmist: ok, I did wonder about that bnep message in the log, didn't think it was serious but seems it was16:05
deathmistye me too, I had to try it tho ;)16:05
deathmistdo system sounds normally not play thru BT audio?16:06
maldeathmist: not sure, I rarely use BT audio16:07
malonly when testing new release16:07
deathmistpretty much last major isse I have is no audio during a call, otherwise everything works pretty much perfectly16:15
malyou are sure you have miniaf running and nothing bad in journalctl from pulseaudio16:23
deathmistmal: miniafservice is running + journalctl
maldeathmist: always provide full log, that is just random part of it, journalctl -b17:14
deathmistmal: there, updated the same paste17:16
maldeathmist: what do you see there when you try to make a call?17:21
piggzsweet, somehow in in a rootfs17:31
Mister_Magisterpiggz: hueh17:33
Mister_Magisterso u fixed it?17:33
piggzMister_Magister: well, im unsure what i did...repeated steps and this time created init_enter_debug2 to halt the boot.....17:35
piggzbooted, got into debug17:35
piggztold it to continue, and im stil here17:35
deathmistmal: here are logs from right before I started a call to after it ended: journalctl logcat dmesg
maldeathmist: a suggestion for future, remove phone number from logs17:37
deathmistgot it :p  guess I searched thru the wrong log beforehand as I didn't see anything17:42
malit was in journal17:43
deathmistyeah  found it, blanked it on the paste already17:43
Mister_Magisterpiggz: feel free to ask me stuff17:48
deathmistoh and audio isn't the only thing not working during a call, same for mic which for sure works outside calls17:51
piggzMister_Magister: sure will!17:53
Mister_Magisteri need to take out trash! it's wet and smells like dead body!17:54
deathmistlmao sending files via BT works on SFOS but it was broken on LOS 15.1 xD18:28
anpepSo I didn't find a way to get the correct fstab entries on hybris-boot. My device tree is bq/bardockpro, which depends on bq/msm8953-common. Under device/bq/msm8953-common/rootdir I was able to find the correct fstab.qcom file, but for some reason the fstab in hybris-boot rootfs is an almost empty one with some default entries, any idea?18:51
anpepThere is no fstab file under device/bq/bardockpro/rootdir18:52
malanpep: what do you mean hybris-boot rootfs? sailfish doesn't use fstab on device at all18:53
malanpep: it doesn't matter if the fstab is in device repo or the common device repo, as long as build finds it18:53
anpepsome days ago I stumbled upon this issue and you suggested looking for fstab files under out/18:53
malI said to look if it contains everything you might need18:54
anpeponly two entries show up, in hybris-boot and hybris-recovery, which contain some default entries but no trace of the correct partitions of my device18:54
mallike system and vendor partitions etc18:54
malwhat .mount services do you have on device in /lib/systemd/system/18:55
anpepactually I didn't get to build the OS image, I started all over again on a clean tree in case I messed up some basic stuff (which I think I did not after all)18:58
anpepI guess I'll keep going and see which .mount units end up on my image18:58
malanpep: you should be getting some fstab in out/ if not then something is wrong19:05
malI assume you are using hybris-15.119:05
malnot sure if I said to you to look for fstab in output of "make modules"19:05
anpepyes, I think I found the fstab.qcom file, but not sure, trying now19:06
anpepyep, fstab.qcom is there19:06
anpepbut not in out/19:07
malif that doesn't not appear in out/ after make hybris-hal then run make fstab.qcom19:07
malalso check for ueventd*.rc in out and output of make modules19:08
anpepfound ueventd.qcom.rc19:08
malin out?19:09
anpepin the output of `make modules`19:09
anpepnot in out/, guess I'll run make individually19:09
anpepI wonder why that does not happen automatically19:09
anpepyep, now I can see them in out/19:11
anpepthanks! But I wonder if there are more targets that are missing from the build, how can I find out?19:12
maltry running make ueventd.qcom.rc and check what appears in out/19:12
anpepalready did, it shows up as ueventd.rc19:13
malthis is a known issue, usually only those two are missing and all should be fine now19:13
anpeplovely, I'll try building the image now, thanks for your help19:13
malanpep: just to be sure, have you checked faq in topic for 15.1 porting notes19:13
anpepthe hadk faq?19:14
malit tells you to copy some files to config sparse19:14
anpepactually I looked for my issues as they were showing up but didn't really read it completely19:14
maljust search for 15.1 and copy the files it tells19:15
malsome .rc files, you'll need those later19:15
anpepok, thanks19:16
malwithout those files you won't get UI working19:16
malthose instructions should be quite clear but if you have questions just ask19:17
anpepthanks! btw my libhybris repository is not the "official" one19:17
deathmistanpep: also in faq (if you didn't do already) downgrade droid-configs-device submodule files, you can just search for ""19:17
malanpep: what is it?19:17
anpeplibhybris was missing from my tree so I looked on the internet and found an IRC log pointing to the android8-initial branch of krnlyng/libhybris19:17
deathmistoh that's old, don't use it19:18
malanpep: DO NOT use android8-initial19:18
maljust clone the master branch as mentioned in faq in 15.1 notes19:18
malI suggest you run make hybris-hal after cloning that correct branch19:19
deathmistbtw why is libhybris still not in the hybris-15.1 default manifest file? adding <project path="external/libhybris" name="mer-hybris/libhybris" revision="master" /> to my local manifest worked perfectly19:21
maldeathmist: because in a way that is more complicated, to update libhybris porter would have to run repo sync19:21
malalso not sure how well repo sync handles submodules19:22
deathmistseems to have fetches everything fine for me, isn't the same being done on 14.1 and older branches too tho?19:22
malolder ones didn't need external/libhybris19:23
anpepadded mer-hybris/libhybris to my manifest and repo did NOT fetch the submodules, even when explicitly told to, so I manually cloned it under external19:23
deathmistanpep: it didn't for me either at first, did a clean build with it there from the get-go and it worked (not worth it lmao)19:23
malalso it doesn't really have to be external/libhybris, also hybris/mw/libhybris works19:24
malanpep: old logs are sometimes dangerous, you can end up using completely obsolete things19:25
anpepguess I learned that the hard way19:26
malfaq is the first place to check, if you don't know if something is right just ask here19:26
T42<DSstill01> Faq->hadk->crying in #sailfishos-porters20:02
T42<DSstill01> That's the order20:03
vknechtshould grep logs before crying, tho20:10
piggzmal: r0kk3rz: gross hack or elegant solution?
piggzignore the fact it isnt a .mount unit20:10
Mister_Magisterfirst one20:10
piggzharsh :D20:11
malpiggz: what are you trying to do in that?20:14
malpiggz: so you didn't want to wait for me to give the bind mount fix for linking20:15
piggzmal: replace the linker .... but it also ensures files loaded from /system are the ones we provide in /usr/libexec...20:15
Mister_Magistermal: i'm still waiting20:15
Mister_Magisteri'm telling him it's bad but he won't listen!20:15
Mister_Magisterbad piggz! you won't get cookies!20:15
piggzmal: sure, i'll get a propoer fix, i was just being impatiant :D20:15
T42<DSstill01> @vknecht [should grep logs before crying, tho], lol20:16
Mister_Magister@DSstill01 don't use reply option20:17
T42<DSstill01> D:20:17
T42<DSstill01> And I?20:37
T42<eugenio_g7> cool stuff 🎉20:38
deathmistmal: you got any fix in mind for in-call audio / mic from the earlier logcat? idk why mentions msm8974_platform when I have a msm8998; some errors also about perf locks, "Invalid snd_device = " and ACDB-LOADER doesn't seem perfectly happy either, here is the full logcat before the call
maldeathmist: check the acdbdata symlink in /etc does it point to something that exists21:09
deathmist"/etc/acdbdata: broken symbolic link to /system/etc/acdbdata" that would explain some things...21:09
malhmm, but it seems logcat says it loads those from /vendor/etc21:10
deathmistguess I could ln -s /vendor/etc/acdbdat /system/etc/acdbdata, don't know if that would help much21:11
malno, if you want to try replace the /etc/acdbdata with a working symlink21:12
malbut as logcat seems to load it anyway it's probably not a problem21:12
maldid you need some changes to somewhere for audio? I mean to get pulseaudio working?21:13
deathmistoh right, I got mixed up with the paths :p  here's the only changes I've done so far to get audio working
deathmistmal: so those changes shouldn't be an issue? probably not related, but after recording with the camera, the output is a bit garbled in the beginning & end and after leaving the app it doesn't launch until a reboot + pulseaudio fully dies21:49
maldeathmist: some devices need a hack to droidmedia for video recording22:02
T42<meierrom> Piggz: Very nice video :) You did it not forgetting about M1M and NotTheKid :)22:11
deathmistmal: where would one find this hack? my SOC is SD 835 and GPU Adreno 54022:25
deathmistI'll try that in a bit (cannot test calling today anymore if that could fix it also), also could I get my cheeseburger fixup-mountpoints modifications upstream as old 14.1 ones already exists for the device but 1+ patches treble in Oreo and made /vendor from unallocated reserve partitions?22:30
deathmistand should they usually be minified (there are around 70 partitions for my device)22:32
deathmistthat also reminded me of this: when I go to Settings > Storage on top of "Mass memory" I also see pretty much all partitions on the phone like /boot /vendor /system etc, could they be caused by this very file?22:34
maldeathmist: that is not related to calls, only to camera22:40
anpepFinally /system partition is correctly mounted, but no UI yet. `droid-hal-init.service` is failing with /sbin/droid-hal-init throwing an obscure selinux error22:58
anpepAlso, the weirdest thing happens when I manually execute droid-hal-init: the device enters fastboot almost immediately for no good reason lol22:58
r0kk3rz@adampigg you broadcasted the secret code!23:07

Generated by 2.17.1 by Marius Gedminas - find it at!