Monday, 2018-07-23

*** ubuntu is now known as Guest5791701:42
*** Guest57917 is now known as zhxt01:46
*** sven_ is now known as _sven04:14
tbrmal: thanks. looks like I broke that when upgrading the underlying OS. Will look into it07:02
tbrmal: looks like it's working again07:04
ohyashdroidmedia is building 7241 targets. the whole hybris-hal seems to have required lesser work than this gstreamer alone07:08
ohyashah, its going faster now :)07:08
ohyashI allowed jolla-settings-networking-multisim pattern. Phone still wont detect second sim.07:19
r0kk3rz[m]repeat after me, droidmedia is not gstreamer07:23
ohyashoh, ok07:24
lbtsteeeve: you aked about an obs repo yesterday - not sure what you wanted.07:43
ohyasherror in chapter 13.2 -> Build relevant parts08:51
ohyashon this command : --droid-hal --mw=
ohyashmaybe r0kk3rz i here to help? :308:53
T4<adampigg> may give the docker based sdk a try09:26
steeevelbt: an obs repo :P09:30
lbtsteeeve: you have a home project by default. You can build packages in there and they'll be published to a repo.09:31
lbtYou probably need to get some guidance on package building etc09:31
steeeveBut i need the one in faq09:32
steeeveTo publish otas09:32
lbtsteeeve: sorry - I'm really busy at the minute. You need to find someone in the community who can help you out. I'd start by being able to clearly state what documentation you're reading, what you need to do and what you've tried and why it doesn't work as you think it should :)09:36
mallbt: I think he probably means he wants a new device to nemo:devel:hw but just hasn't bothered mentioned for which device09:37
lbtmal: sure. I tend only to do that if sledges or you give me the OK that the user is ready to manage a devel project anyway.09:40
steeeveyes the nemo:devel:hw:samsung:a5ultexx09:40
malsteeeve: sledges is usually the one adding the new devices, lbt then patternizes those later09:41
steeevesledges: ping ^^09:41
lbtmal: that's the problem when you become a responsible dev - you start getting responsibilities ... if you check out he's done it all in his home project and you're happy then sledges and I would be happy with that09:42
steeeveI didnt do anything in home repo :/09:45
lbtsteeeve: you usually get stuff working in your own home area - then others can help you get it ready to publish to the community devel area09:46
steeeveIll start doing something there, then09:48
lbtnp - the only difference is the name. Think of it as your personal development space09:48
steeeveSeems legit09:48
steeeveWhy cant i add additional path to my repo?10:02
malof course you can10:06
steeeveNo i cant10:09
steeeveIf i click repository tab i can select anything10:10
malwell I know nothing of that gui way of adding dependency repos, I always add the manual way like here
steeeveOh yeah, ill try10:17
steeeveI cant save the new file10:20
malif you can't save the file it means you are doing something wrong10:23
malbecause it won't save it if the syntax is wrong10:23
steeeveNo, it's correct, ive only added the 2 lines about the path10:25
malshow the file you are trying to add10:25
malsteeeve: you didn't understand what I meant, you only add the common repo there, obviously, there is no nemo:devel:hw:samsung:a5ultexx yet10:28
malyou can't add non-existent dependencies10:28
steeeveAhh, i think i understand10:29
malsteeeve: faq only said about adding common repo so you should have just done what it said10:31
steeeveI created a subproject for hw:common and it doesnt work10:38
malsteeeve: why would you need a fork of common?10:39
malthe point of common is that it's common10:40
malyou are now misunderstanding everything10:40
steeeveNow i go to work ill be back later10:41
ohyash_whats ofono?11:21
malthe middleware handling modems, so calls, sms and mobile data connection11:25
ohyash_mal: So faq->"Ofono problems" is the thing to be followed for getting second sim to show up right?11:28
neothefoxhey again everyone, I can't install bluez5-libs-devel to a sb2 target, and I need them to build rfkill
malohyash_: for second sim instead modifying /etc/ofono/ril_subscription.conf you should instead add a file like this
malohyash_: of course name the file to something better11:30
malneothefox: does your device have new enough kernel for bluez5?11:31
ohyash_mal: any name would work?11:31
malohyash_: yes11:31
neothefoxmal, 3.4.67, how is that?11:32
malneothefox: not going work with bluez5 unless you use backported bluetooth drivers in your kernel11:32
malneothefox: if you want to do that then faq has instructions11:33
neothefoxmal, ah, so I should replace it with regular bluez, including the patterns?11:33
neothefoxmal, my bt is broadcom, and I don't have a more recent kernel for the device anyway11:34
malif you have normal 3.4 kernel then you need to use bluez411:34
malneothefox: it might be possible to also use backported drivers on broadcom but nobody has tried it yet, so better for now use bluez4 and get bluetooth working and then maybe think of bluez511:35
neothefoxmal, ok, and while I am at modifying the patterns what would be the right way to solve /etc/ofono/ril_subscription.conf conflict between ofono and my config?11:37
ohyash_mal: where would be the bluez instructions for kernel 3.10?11:52
malafaik 3.10 should support bluez5, but sometimes backported driver work better so if you have problems then you can try those11:53
ohyash_hadk has nothing written about bluetooth. and faq only has the backport instrustions. Does that mean only installing bluez5 is enough? (Pattern is there already it seems)11:56
malohyash_: depends on which chip your device has11:56
malohyash_: faq lines 479->11:57
ohyash_oh ok, thanks11:58
neothefoxso, my kernel has CONFIG_FIRMWARE_IN_KERNEL, and yet there are no ttyHS* devices to be found12:25
neothefoxthe rfkill device for the bluetooth is there12:26
malyou might need to start the bluetooth subsystem somehow, usually there is some /sys path to start it or writing something to correct /dev device12:27
neothefoxmal, hciattach /dev/ttySAC0 bcm43xx worked, nice12:34
ohyashmal: How do I fix performance issues? The animations are laggy.12:37
malohyash: faq, again12:37
neothefoxsound works with bluetooth12:40
*** jbadiapa_ is now known as jbadiapa12:55
neothefoxok, so to run hciattach at boot should I make a systemd unit or what?13:10
neothefoxthe way described in FAQ doesn't work with bluez4 I guess because I've tried13:10
malneothefox: it should work, maybe you missed something13:13
neothefoxthe ofono part, do I need it?13:15
ohyashneothefox, for dual sim, yes I guess13:17
neothefoxohyash, it's a single sim device13:18
ohyashsingle sim on mine works fine woth any ogono work13:18
ohyashmal, conflicts during mic! others) in both droid-hal-$DEVICE-xxx.armv7hl and droidmedia-xxx.armv7hl13:19
malohyash: you always need to package droidmedia according to instructions if you build the libs before packaging droid-hal13:21
malohyash: so you build droidmedia stuff in HABUILD_SDK but didn't package those (pack script and package build) before you tried to package droid-hal using (with -d or running the whole thing)13:24
neothefoxmal, yeah, see, this is done via service
malneothefox: yep, both methods are possible13:25
ohyash_the --build=hybris/mw/droidmedia-localbuild thing you mean?13:27
ohyash_I did that13:27
ohyash_or I'll redo it now. Is the sequence :: make *droidmedia* -> build_packages --droidmedia-localbuild -> build-packages *all*13:28
malbut did you run before doing that? you need to run the whole first part of 13.2 always if you build droidmedia in habuild13:28
ohyash_I copy pasted them all one by one. Very less chances but maybe I've missed. Doing it.13:29
malohyash_: I have a text file where I have gathered all build steps, I just copy those from there when I need them13:34
ohyash_mal: Good idea. or maybe not even required to copy. Set build environments and then just "./" :D13:37
malI have been too lazy to make a proper script, I have been planning on making automated build system for the android side stuff13:38
ohyash_including the "download everything" part?13:39
maland only for personal use13:39
ohyash_Oh. and I though hadk is going to be real small soon xD13:40
ohyash_but once I get my thing running well. Imma write an article and a script to "build for tomato"!13:41
ohyash_I have to check the build from obs thing too. Seems better than building all day on my dual core system13:41
pseudodevohyash_: what device?13:43
pseudodevoh wait. its tomato13:44
ohyash_YU Yureka (tomato)13:44
ohyash_ah yep13:44
pseudodevnever thought a device would have the codeword tomato13:44
ohyash_pseudodev, Today I came to know about a cat named malloc13:45
ohyash_mal: how do I pack_source for ofono?13:46
pseudodevohyash_: was the cat programmed in C?13:47
malohyash_: wtf are you talking about13:47
malohyash_: pack source way is only used for packaging android side stuff for sailfish13:48
ohyash_mal: the last warning
malohyash_: rtfm, faq once again has the fix for that13:48
ohyash_pseudodev, the owner seems to like C a lot. So maybe. xD13:48
pseudodevohyash_: sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -m sdk-install -R zypper in droid-config-tomato13:49
pseudodevi think that would do it13:50
malthat no way to do anything13:50
malyou can't have two packages containing the same file13:50
neothefoxok, now with bluetooth out of the way the two remaining issues I can find are GPS and Audio14:02
neothefoxas I understand it the sink and the outputs are detected, the codecs do work, but somewhere along the line the sound can't go to the hardware14:03
malgps usually just works, did you remember to enable it in settings?14:04
neothefoxmal, you mean in system settings on device? yup14:05
neothefoxbut test_gps segfaults and apps can only use cellular positioning14:05
malwell then again logs when trying to use gps, soon after reboot like usual14:07
neothefoxmal, ok, give me a second14:08
neothefoxmal, journal - kernel -
neothefox and this is test_gps output14:13
neothefoxoh, also I've run test_audio, and it might be relevant14:13
neothefoxAudio device API version 0200 doesn't match platform current 3004.test_audio: test_audio.c:71: main: Assertion `audiohw->common.version == ((((3) & 0xff) << 8) | ((0) & 0xff))' failed.14:13
mallogcat is the important log when things have something to do with android like in case of gps14:14
maltest_audio is not a good test, it usually doesn't work14:14
malneothefox: that jounral log is useless, it has only a tiny snippet of the whole log14:15
neothefoxmal, logcat, let me get the journal one more time14:16
neothefoxhuh, note the audio_hw_primary: Amplifier initialization failed14:16
malneothefox: pastebin output of ls -l /etc and ls -l /system/etc14:17
neothefoxmal, etc -
malneothefox: maybe try adding something like this (that is a symlink) check if you have some file with similar name in /system/etc/14:19
neothefoxmal, /system/etc -
malignore that last one from me, you don't have such file14:20
malneothefox: on device run:  ls -l /system/etc/mixer_paths_0.xml /etc/mixer_paths_0.xml14:21
malneothefox: oops it should be ln -s /system/etc/mixer_paths_0.xml /etc/mixer_paths_0.xml14:22
ohyash_how do I know my hardware chips and versions of device?14:24
ohyash_like lspci on system14:24
neothefoxmal, nope, did nothing14:24
neothefoxohyash_, you can look at /sys/class catalog, and then look for your devices by finding "name" file14:31
ohyash_neothefox, bluetooth folder is empty :/14:34
neothefoxohyash_, as mel pointed out you should use hciattach to try and attach your device via one of /dev/tty* devices14:34
neothefoxohyash_, try looking at  /sys/class/rfkill devices, they might give you a clue about what bt device you should be looking for14:35
ohyash_cant hciattach over bluetooth. The bluetooth page in settings is pure blank and bluetooth wont turn on14:43
ohyash_and rfkill folder is vlank14:43
ohyash_and rfkill folder is blank14:43
neothefoxohyash_, how do you do hciattach?14:44
*** doniks_ is now known as doniks15:04
neothefoxI fixed the sound!15:16
neothefoxall it took was to re-flash LOS base and then flash back my build, no idea why that worked15:17
neothefoxso now just a few finishing touches and I am ready to publish this15:17
malneothefox: did you fix gps also?15:25
neothefoxmal, nope, that is still where it was15:25
neothefoxbut I think I'll publish it as soon as I'll make it switch to headphones and back.15:26
steeeveIm trying to look at webhooks examples, however it requires login. Why? How can i see theem?15:30
ohyashwhy is browser performance so worse compared to android?15:31
neothefoxohyash, have you built the gstreamer package?15:33
ohyashneothefox, yes15:33
neothefoxmal, welp, one reboot later it went back to not working15:33
ohyashcamera and phone calls working. so I guess its a yes15:33
T4<adampigg> finally, figured a solution for my blocking kde-connect review!16:21
T4<adampigg> next, id better rebase it16:22
wiktorek140mal: ping16:35
malwiktorek140: ?16:35
wiktorek140mal:  some troubles with audio., gstreamer work, video work, mp3 and other work. audioflingerglue installed, but no sound during calls16:37
malwiktorek140: and the other part of audioflingerglue i.e. the pulseaudio plugin?16:37
malwiktorek140: which device?16:37
wiktorek140mal:  surnia, devicetree not updated yet16:38
malonce you have the stuff in github give me a link to your config repo16:39
wiktorek140pulseaudio, pulseaudio-modules-droid/common/glue16:39
wiktorek140mal:  sure :)16:39
malwiktorek140: did you build things on OBS?16:40
wiktorek140mal: not yet16:40
wiktorek140mal: hybris part is updated (
wiktorek140mal:  of course branch hybris-14.116:57
wiktorek140mal: if i forgot tu update something, please point it to me17:04
neothefoxdamn, I can't even reproduce this working sound17:29
neothefoxcould be a race condition then17:29
neothefoxmel, do you know what should be in  /etc/audio_platform_info.xml? It's a broken link17:38
neothefoxthere is no such file in /system/etc17:38
malneothefox: ignore it17:41
malwiktorek140: pastebin ls -l /etc/ and ls -l /system/etc/17:43
novacomdMy phone stucks at vendor logo. Logs from Sailfish OS . LineageOS 14.1, Android 7.1.217:47
pseudodevwiktorek140: Are you doing for athene?17:51
wiktorek140pseudodev: nope, surnia17:51
pseudodevOh. I did for otus17:51
pseudodevthe 3g version of the device you are working on17:52
wiktorek140pseudodev: i was trying previously build one for otus but missing device make it hard ;)17:52
pseudodevWhy are you doing with CM14?17:52
wiktorek140because i can, and cm13 and cm12 was troublesome17:53
pseudodevAs far as i know it (maybe things have changed now), but older versions of cm have better support17:53
pseudodevwhat kind of trouble?17:53
wiktorek140pseudodev: cannot fully reboot17:53
pseudodevBut the build booted up right?17:54
wiktorek140on cm14 is the same. but it was easier with audio, wifi and bluetooth17:54
wiktorek140pseudodev: yes, on all cm12/13/14 it booting to gui17:54
pseudodevSo maintain 1217:55
pseudodevit will be more stable17:55
wiktorek140nope, it has the same bug. dont reboot17:56
wiktorek14014 is more stable for now17:56
wiktorek14012 often crash transmission17:56
malnovacomd: you probably didn't run the symlink script mentioned in 14.1 porting section of faq (linked in channel topic)17:57
novacomdok. I'll try. Thanks17:58
wiktorek140mal: did you know which patch is required when require it?18:16
pseudodevno rule to make drivers/bluetooth/hci_smd.o18:34
wiktorek140probably missing line in Makefile (
wiktorek140pseudodev: or use this commit ( and compare with that u have18:40
pseudodevalready found it. but thanks :)18:41
pseudodevHuh. seems like i dont have the bluetooth module itself.18:53
pseudodevWhat module do i need to build for msm8952?18:54
pseudodevNote: i already have CONFIG_BT_HCISMD=y18:54
r0kk3rzwhat bt module do you have?18:59
neothefoxpseudodev, there is a good chance you don't even have modules19:05
neothefoxpseudodev, which should not be a problem, you probably have to activate the device with hciattach19:06
pseudodevhow do i do that?19:06
neothefoxand it could be any /dev/tty<letter><number> device19:06
r0kk3rzfind out what the thing is first19:06
pseudodevr0kk3rz: How do i find that?19:07
r0kk3rzbluedroid configs, boardconfig.mk19:07
neothefoxpseudodev, hcitool -l would show you the list of possible chips, select the one you have19:07
neothefoxpseudodev, and then it should be like hciattach /dev/ttySAC0 bcm43xx19:08
neothefoxonly replace the tty and chip name19:08
neothefoxalso you would need a firmware which you must name as your full bt device name (it would show up while running hciattach)19:09
neothefox I should make it a symlink, but you should get the point19:10
r0kk3rzsettle, it might not be a broadcom one :P19:10
pseudodevIt is athene19:12
pseudodevso it has snapdragon19:12
r0kk3rzthat doesnt mean shit19:13
r0kk3rzfind out which bt it has19:13
r0kk3rzits qcom at least, so you might need to port in the hci_smd driver19:19
neothefoxby the way, is it ok for a few mounts such as /config and /cpuctl to fail at boot?19:22
pseudodevr0kk3rz: I think it has HCI_SMD19:25
pseudodevi did everthing mentioned in the faq19:25
*** Nokius_ is now known as Nokius19:26
pseudodevand it doesnt woork19:30
r0kk3rz'it doesnt work' wow great analysis there19:34
r0kk3rzvery interesting19:34
neothefoxpseudodev, do you have /sys/class/rfkill devices?19:35
neothefoxpseudodev, try looking at their names, do you have any bt ones?19:44
ohyash_web performance is literally terrible. any workarounds?20:12
neothefoxohyash_, can you confirm that hardware acceleration works?20:15
ohyash_neothefox, the test_hwcomposer thing?20:16
neothefoxohyash_, well try that, but I don't really know if that works at the moment20:17
ohyash_I'm getting the spiral20:19
r0kk3rzhuh? of course it works :P you just need to use it right20:19
ohyash_I got gstreamer. Videos do play. Calls work.20:20
ohyash_I've done the graphics improvements thing from faq as well. Performance is still not good. I can feel the lag in the OS. and web browsing is worse.20:21
ohyash_Native apps lag too. (The slide transition animations and scrolling when there are images)20:22
ohyash_Oh, btw, accelerometer sensor isnt working either20:23
malohyash_: did you add the performance trick from faq?20:27
malohyash_: also did you add the sensor fix from faq20:27
ohyash_mal: performance trick. check. It helped in a getting a little bit smoother. But still is laggy if I compare it with android performance on the same device. It feels like I moved from a compiled ecosystem to interpreted20:29
malohyash_: does the device have two types of cpu cores?20:36
ohyash_I guess yes20:36
malsome of those have issues because the smaller ones get used instead of the big ones20:36
ohyash_and nope, havent done any sensor fix. I searched for "accelerometer" in faq and that wasnt helpful. I didn't much, will revisit20:37
malohyash_: search for sensor20:37
ohyash_only smaller cores working could be a problem.20:38
guhlhi mal, hows life?20:43
malbad as usual20:43
guhli am trying to reappear after some very hard month (work related)20:44
guhlI have seen you did some commit to the 15.1 branch of libhybris20:45
ohyash_mal, if you had an option to either improve your life or my web performance on sailfish what would you choose? :320:46
guhldid you do a working port on 15.1 or is this still experimental?20:46
malohyash_: you think me helping here makes my life better?20:46
malguhl: libhybris is working fine for 15.1, there are already two hybris-15.1 based test ports from here20:47
malguhl: I did the early work for the new libhybris linker and krnlyng finished it20:48
guhlcool, my state is: I have a stable self built LOS 15.1 for the chiron (Xiaomi Mi Mix 2) and I started with the hybris built20:48
malguhl: note that modem stuff won't yet work in android 8 based ports20:49
guhlmissing some dependencies atm but I'll figure that20:49
ohyash_mal: certainly not. I see you here correcting hideous mistakes of me and other similar porters. It clearly never improves. but I was just trying to make a joke, (and catch some help meanwhile xD)20:50
guhlmal, good to know - did not work on my experimental 14.1 port either :-)20:50
T4<adampigg> guhl, welcome back20:50
malguhl: it needs libhybris cloned to for example external (the branch from the libhybris PR)20:51
malohyash_: well today has been an unusually bad day for me so nothing is funny today20:52
ohyash_Oh.. ok.20:53
guhlmal, could you elaborate this a bit20:53
ohyash_don't you have alternate maintainers who'd handle us while you take rest.20:54
malguhl: ?20:55
guhli don't understand that sentence "it needs libhybris cloned to for example external (the branch from the libhybris PR)"20:56
guhlwaht is "the libhybris PR"20:56
malguhl: in android 8 you have to have libhybris somewhere where android build finds it, like under external or hybris folders20:57
* guhl did only to finance work the last 8 month so i might be a little bit slow20:57
malguhl: if you look into upstream libhybris PR list you will see an android 8 PR20:57
guhlmal, OK got it20:58
neothefoxI think I am on to something with the sound. strace shows that the system looks for and for, both are not in any path, and they are not in the LOS build too21:30
neothefoxI've tried symlinking them to /system/lib/hw/ and /system/lib/hw/, it tries to load them but fails21:31
malare there any audio_amplifier libs on the device?21:31
malalso sometimes it might look for some extra libs that are not required21:32
neothefoxmal, nope, and since it's not in the LOS build I don't think they are really needed or even exist, however could that mean that hybris just wrongly configured?21:33
malno, it's possible it always looks for those, even in android21:33
malbut you said it worked on first boot? one option is some timing issue?21:34
neothefoxmal, that worked when I've flashed the android base again before flashing the Sailfish, but I can't reproduce it anymore. So that hints at the issue being not with the files, yup. But I have no idea what and how can I delay during boot21:35
malneothefox: have you tried restarting pulseaudio? does it help?21:36
malsystemctl-user restart pulseaudio (or systemctl --user depending on which user you are)21:37
neothefoxmal, no, restarting does nothing (I've tried pkilling it, and I've just tried restarting with systemd)21:39
malwhat is that?21:39
neothefoxhere is the audio_test strace (I used it because it shows system loading audio stuff, not because I rely on audio_test to be viable)21:39
mallike I said earliet test_audio is useless21:40
malif you want to strace something strace pulseaudio21:41
neothefoxok, give me a moment21:41
malstrace attach is useful21:41
malpreferrablu after reboot so everything is seen21:41

Generated by 2.17.1 by Marius Gedminas - find it at!