Sunday, 2018-09-16

*** ChanServ sets mode: +v T405:37
*** n-nani is now known as Guest1347408:40
Mister_Magistermal: can i make update?10:54
malworked without issues for me, I assume you have latest config submodule and other packages11:01
Mister_Magisternah i dont11:01
Mister_Magisterso update submodules and try? kk11:01
malthat's the usual way11:02
r0kk3rzyeah seems ok here11:04
Mister_Magisterkk thanks11:05
malconfig submodule is the most important usually11:08
Mister_Magisteryea i know11:15
T4minlexx was added by: minlexx11:16
Mister_Magistermal: what was the version again? 2.2.1.?11:27
mal2.2.1.18, you do know that there are release notes on tjc which would have told you that11:30
Mister_Magisteryee i already looked at sailfishos project at obs sorry i shouldnt ask11:31
T4<abhishek_0> Do we have a fix for colors issue yet ?12:23
r0kk3rz@abhishek_0: have you fixed it?12:24
T4<abhishek_0> you mean taking the dive into gst's code ? no12:29
T4<abhishek_0> XD12:30
r0kk3rzwell, who do you expect is doing that?12:31
elros1from my little understanding how it works you could at least try to use code from QOMX_COLOR_FormatYUV420PackedSemiPlanar32m case12:31
elros1as a base12:32
T4<abhishek_0> let me try12:34
T4<abhishek_0> ?12:35
r0kk3rz@abhishek_0: did it work before?12:41
T4<abhishek_0> nah , i had to use omx, i just got rid of omx12:42
T4<abhishek_0> few others have the same issue, piggz confirmed it12:43
r0kk3rzyeah then that section of code needs filling out for whatever colour format it is12:44
Mister_Magistermal: halp
malsomeone had the same issue before13:10
Mister_Magistermal: do you remember fix?13:11
T4<abhishek_0> ah it is also on my gitlab-ci
T4<abhishek_0> due to outdated targets13:15
ghosalmartintime to build another 50k worth of objects. typing make clean as a habit is a bad habit :P13:15
T4<abhishek_0> i dont even close the same terminal ,, because it will wive the source environments13:16
merbotMer bug 1956 in mic "Mic crashes when rpm package contains non numeric characters other characters." [Normal,New]13:19
abransonmal: that should have been fixed here:
abransonwhich should be in 2.2.113:26
T4<adampigg> Abranson: any thoughts.on color conversion issues?  anything i can supply?13:30
malMister_Magister: so update you platform sdk (not the target), the usual way with zypper13:38
maland ssu13:38
malabranson: ok, so then the issue is that people forget to update platform sdk13:39
r0kk3rzMister_Magister: see, i told you to make sure you update your sdk :P13:42
malr0kk3rz: he probably only updated the target13:43
malin the past updating sdk hasn't been critical13:43
r0kk3rzi know13:43
r0kk3rzthats why a specifically mentioned it13:43
r0kk3rzbecause this time, it is critical13:43
malprobably Thaodan had the same cause for his issues13:47
Thaodanyes, I fixed it on my own at that point but abranson said it got pushed to 2.2.113:48
malThaodan: maybe close the bug since it was fixed already even without the patch in that13:49
Thaodandone just now13:50
T4<abhishek_0> r0kk3rz will you be updating your docker image ?13:52
ThaodanI would be great if someone could help me debugging me mmc issue13:57
Mister_Magistermal: not found :/13:57
Mister_Magistermal: i didnt update target13:57
Mister_Magisterobs is my target xD13:58
malMister_Magister: seems like an old path13:58
Mister_Magistercan you give me your ssu lr from sdk please?13:58
malI have adaptation0 ...
Mister_Magistermal: i have this as adaptation114:00
malMister_Magister: so delete the incorrect one14:01
Mister_Magisterhow do i delete global repo14:01
malor disable14:01
Mister_Magisterye disabled it14:01
Thaodanedit ssu.ini14:01
Mister_Magisterdup or up14:02
Mister_Magisterversion --dup14:02
Thaodanzypper dup is distro upgrade14:02
Thaodanup is regular14:02
Mister_Magisteri know14:03
* Mister_Magister hopes sdk wont break14:03
malThaodan: wondering how the mmc is connected in the device14:05
Thaodanfcc released the shema of the xa by accident14:06
Thaodancan I get more debug information besides changing printk to 8?14:08
abransonmal, r0kk3rz: it'll need to be done for the next update too - that'll have the zypp updates14:09
abransonadampigg: which colour conversion issues?14:10
r0kk3rzabranson: it would appear there is some colour format 21 that needs implementing14:11
abransonmal: for browser video? which device?14:12
malabranson: fp214:12
T4<abhishek_0> redmi 214:12
T4<abhishek_0> and modo14:12
T4<abhishek_0> *mido14:13
T4<BirdZhang> redmi 5 plus14:13
abransonfor browser video playback?14:13
T4<BirdZhang> yeah14:13
T4<abhishek_0> specifically with sailfish-browser14:14
malabranson: I have tried the quirks, use-codec-supplied-height is required for video to play at all, use-codec-supplied-width crashes browser on video playback whether or not use-codec-supplied-height is defined14:15
malabranson: I assume this is correct way to define both14:16
malvideo/avc = use-codec-supplied-height14:16
malvideo/avc = use-codec-supplied-width14:16
abransonin the new sdk I think there's some new stuff in sdk-assitant to update tooling and targets together14:16
r0kk3rzabranson: that didnt work for me, so i did it manually14:16
abransonbah androidxref is down14:19
abransondoes anyone know what colour format 21 is?14:19
elros1when I briefly search it yesterday I think it was OMX_COLOR_FormatYUV420SemiPlanar14:21
abransonthat's a qcom one, has a much higher number. 7Fxxxxx or something14:22
elros1but I could be wrong14:22
abransonone that low should be quite standard hopefully14:22
Mister_Magistermal: whoa mic got updated?14:23
abransonelros1, mal: that looks like HAL_PIXEL_FORMAT_YCbCr_420_SP, which might work as I420 straight.14:25
malabranson: afaik my issue is not related to color formats14:26
malMister_Magister: yes14:27
abransonmal: sorry, though you had this too14:27
malabranson: my issue is just browser crashing when seeking youtube videos14:27
elros1@abhishek_0 have it14:28
abransonin which case, adding 0x21 as another option to should fix it14:28
T4<abhishek_0> yes, want logs ?14:28
abransonmal: not sure about the crashing on seek sorry. some logs might help there.14:30
T4<abhishek_0> here is the log of sailfish-browser with this library present ,
elros1think it was 21 decimal14:30
T4<abhishek_0> and without that lib
abransonabhishek_0: and what does the video look like for each of those?14:33
T4<abhishek_0> without , no video at all , only audio comes out ,,, with weird colors issue, they are distorted14:35
abransonhmm that's odd - with the .so you should get the right colours. is it definitely the correct lib for that device?14:36
T4<abhishek_0> yes, same issue piggz is having14:36
T4<abhishek_0> and maybe birdzhang14:37
T4<BirdZhang> yeah, the same14:37
abransonhmm, the mozilla source says that one should correspond to OMX_QCOM_COLOR_FormatYUV420PackedSemiPlanar64x32Tile2m8ka14:38
malabranson: which logs do you want, I can give backtrace but that didn't seem very useful14:38
mal#0  0xb617681a in __pthread_mutex_unlock_usercnt (decr=1, mutex=<optimized out>) at pthread_mutex_unlock.c:4114:38
mal#1  __GI___pthread_mutex_unlock (mutex=0xe5e5e5e5) at pthread_mutex_unlock.c:31014:38
mal#2  0x96f45684 in ?? ()14:38
abransonah no that's the tiled variant14:38
abransoncolour formats are weird14:38
malabranson: logcat only has14:38
mal09-16 14:36:33.456 21339 21518 E DroidMediaCodec: Got EOS14:38
mal09-16 14:36:33.452 21528 21528 W sailfish-browser: type=1701 audit(0.0:75): auid=100000 uid=100000 gid=100000 ses=2 comm=4D65646961204465636F6465202333 reason="memory violation" sig=1114:38
abransonmal: gst_debug?14:39
malabranson: this shows some of the normal playback loops and then what happens after seeking
abransonmal: hmm that is weird, especially after seeking14:56
T4chronosmsx was added by: chronosmsx14:59
malabranson: any ideas how to debug more15:00
T4<chronosmsx> #whois … Location: Russia … Name: Michael … Age: 19 … Task: Porting sfos to Galaxy Note 3 exynos15:00
malabranson: maybe I should test again if removing some reverts would help, I had to revert some changes in frameworks/av because metadata spoofing fix in android15:02
wolfquinhello, i wonder where i could find the branch of a old Samsung device by the model sm-t560, since the manifest that i had to copy in .repo/local_manifest doesn't mentiones it, and also the manifest in the .repo folder. anyone got any clue?15:11
wolfquini am porting an unofficial LineageOS so i can't find it on their wiki15:12
malabranson: I think that metadata issue is only in android 6 base, droidmedia with android 7 uses a different way15:12
ghosalmartinis anyone working on any new gen devices? made in the last year or two?15:29
malghosalmartin: why do you ask?15:32
ghosalmartincompletely out of ideas as to why CONFIG_VT is breaking, so curious if anyone else had similar symptoms...any has magically fixed the problem :P15:33
Thaodancan I get my kernel be more verbose other than setting printk to 8=15:35
malghosalmartin: what was 4.4 kernel?15:35
ghosalmartinmal: chiron?15:36
malwell what ever device are talking about15:36
ghosalmartinmi mix 315:36
malghosalmartin: so the issue is in some 4.x kernel15:37
ghosalmartinmal: yes it applies to my Pixel XL as well, same issue15:37
malxa2 has 4.4 kernel and no CONFIG_VT issues15:37
elros1Thaodan: sometimes there are sysfs attribute which increase log level for particular driver/subsystem or it can be increased in sources15:38
ghosalmartinmal, feels like because its using AOSP as base15:38
malghosalmartin: wondering if there are any working lineage based ports with 4.4 kernel15:40
ghosalmartinsince when i use xiaomis original kernel it doesnt reboot immediately but telnet doesnt work and no ramoops15:40
ghosalmartinmal, the lineage port works perfectly but it doesnt enable config_vt15:41
malghosalmartin: a patches xiaomi kernel15:41
malghosalmartin: init-script is a bit picky about configfs drivers, I can only get telnet working in there when using certain drivers15:42
ghosalmartinmal: yeah but it would leave a lovely consolerammops15:43
T4leoaslan was added by: leoaslan15:44
malghosalmartin: ?15:44
ghosalmartinmal: id assume the reboot is a panic, and if it panics it would leave logs to recover15:44
malhow soon does it reboot?15:45
ghosalmartinnot instantly like the config_vt breaj15:46
ghosalmartinso that is possibly fixed15:46
maldoes host dmesg show anything15:48
ghosalmartingoing for another build so ill let ya know15:50
ghosalmartinmal the dmesg is clean :(16:01
wolfquinokay so i've tried to google what a branch is and appreantly its the what version of the source code it is? can anyone explain to me if its wrong? and sorry if it sounds like im spamming this chat.16:03
Thaodanelros1: thanks16:04
Thaodanmal: can you give me dmesg output of your xa2 were sdcard support is working?16:05
malghosalmartin: what do you mean by clean? no sign of the device? no errors? normal mer boot messages?16:13
ghosalmartinmal: no sign of device at all16:14
ghosalmartini might try and build the xiaomi kernel and see if I can just get it to boot android first16:14
T4<meierrom> Mal: have you successfully ported sfos to xa2?16:17
r0kk3rzi think i saw someone trying to build the xa2 adaptation16:24
T4<meierrom> @r0kk3rz [i think i saw someone trying to build the xa2 …], Yeah, me too. So mal didn't port it yet?16:30
r0kk3rzno idea16:35
r0kk3rzhow far did you get building it?16:35
T4<meierrom> Nowhere. Just researching :)16:37
Thaodanr0kk3rz: I adapted the adaptation to the xa2 ultra16:52
Thaodanbluetooth and gps is not working16:53
Thaodansdcard for my ultra16:53
Thaodanbesides that it works fine16:53
malThaodan: there is gps support for android 8 devices waiting to bo merged, the PR is in relevant mer-hybris repo16:55
Thaodanmal: ok than I was out of date16:57
malit requires building using different spec file than the one used for older adaptations16:58
ThaodanCan you give me dmesg output when booting of your xa2 with working sdcard ?16:59
r0kk3rzThaodan: oh nice! same process as xperia x?17:07
Thaodanr0kk3rz: no but same kernel17:07
Thaodanor do you mean building process?17:07
r0kk3rzbuilding process17:07
Thaodanneeds some update packages from -devel17:08
r0kk3rzare those in 2.2.1? or still in future release17:09
r0kk3rzah ok, yeah configfs stuff and binderised ofono17:09
Thaodannew bluez is needed to17:36
Thaodanneeds support rome based chips17:36
malthere is no support for rome chips in bluez afaik17:36
Thaodanthere is but for 5.1917:37
Thaodanmade by qualcomm and not merged17:37
ThaodanI rebased the tree and need to fix some light compile errors. Sadly I was sick the weekend and don't had the time17:37
malThaodan: does it work?17:38
*** ChanServ sets mode: +v T417:38
malI also found some patches but nothing that works17:38
Thaodanmal: as I said didn't had the time to test further cause I was sick17:39
malThaodan: also xa2 doesn't have a rome chip but a related cherokee chip
Thaodanok. than there's still something missing when qualcom contributes mainline support for that chip17:42
ThaodanI mean makes mainline support without bluez sense?17:45
malThaodan: hmm, those look a bit different than the ones I found, I'll have a look17:49
entilmal: I can't hang out right now, but I ran out of battery and on starting the phone again, I wasn't asked for a pin code, but mobile connectivity works - is there some kind of pin code cache now?18:05
malentil: not that I know of18:06
entilmal: kk, will fiilistele this later, thx, ttyl!18:06
piggzanything porters can do for fingerprint sensors yet?18:07
malpiggz: I think not18:08
wolfquinhello, anyone knows what is the cause of the error; error: Exited sync due to fetch errors18:13
wolfquinit always gives me the error at 19%, i think its because something is wrong with my manifest, tho one of my friends says it looks correct.18:14
malwolfquin: pastebin your local manifest and/or pastebin the output of repo sync18:34
r0kk3rzwolfquin: try sync without fetch-submodules18:35
piggzmal: abranson: fscking around with gst-droid ... i can get video to play after removing ... but the colour are not correct yet19:59
piggzwith the lib there, gst-droid takes the path20:00
piggzif (use_droid_convert) {20:00
piggzwhich i guess uses the android side20:00
piggzremoving the lib will make it use an inernal path20:00
abransonpiggz: that is really odd - that should be the native colour conversion library.20:11
piggzabranson: well, it doenst work, and crashes the browser :)20:12
abransonMaybe the interface was changed or something20:12
abransonWe always used that lib until the sony, which was the first device that didn't have one20:13
malpiggz: it directly crashes?20:16
malpiggz: after starting video playback?20:16
r0kk3rzmaybe its not ordinarily used on android?20:17
malpiggz: are you sure your codec quirks are correct if video playback crashes browser?20:20
T4<abhishek_0> with library it segfaults for me ,, as per gst_debug20:22
piggzmal: yes, it crashes after starting playback ... what are the quirks?20:23
malpiggz: there is also one for width20:24
ghosalmartinanyone know where pulseaudio is provided from? Warning: repo problem: nothing provides pulseaudio >= 11.1+git4 needed by pulseaudio-modules-droid-11.1.73-1.1.4.armv7hl,20:31
Mister_Magisterupdate it on obs?20:31
malghosalmartin: do you have correct target?20:32
piggzmal: height, and height + width shows no difference20:32
Mister_Magisteror in local repo depends on where you build it20:32
ghosalmartinmal: am running off obs20:32
ghosalmartinso not sure if the target matters?20:32
malpiggz: try the last option width-only20:32
Mister_Magisterghosalmartin: link?20:32
malghosalmartin: link to your project20:32
Mister_Magisternot found lol20:32
ghosalmartinoops :P
Mister_Magisterseems ok to me20:33
malghosalmartin: you have wrong common target20:33
ghosalmartinmal: ahh is it meant to be latest?20:33
piggzmal: nope, crashed again20:33
Mister_Magisterah ye20:34
malghosalmartin: I meant wrong normal target20:34
malghosalmartin: check here
piggzmal: hang on, it hasnt pronted the debug messages which say it used the quirk...20:34
ghosalmartinmal, jsut did thanks :) thats my fav OBS project :20:35
* Mister_Magister feels that he cant help20:35
malghosalmartin: checking my github and OBS (or fp2 OBS) are usually good sources for help :P20:35
* Mister_Magister help people*20:36
Mister_Magisteryee mal is good and stuff20:36
ghosalmartinmal: has it changed recently cause i built a zip a couple weeks back with no issues20:36
malMister_Magister: it's easy to miss the incorrect version in targets20:36
ghosalmartinaltohugh i think i had pulseaudio in the project20:36
malghosalmartin: latest is now
Mister_Magistermal: ye i just thought when i try to help too and im trying my best helping i would get some "Thanks" too :/20:36
ghosalmartinmal: ahh awesome wasnt sure if it was added20:36
malghosalmartin: the problem was the mismatch between normal target and common target20:37
ghosalmartinmal: hmmm same issue atm20:39
malghosalmartin: you are missing droidmedia and audioflingerglue rpms from droid-hal-chiron20:40
Mister_Magisterghosalmartin: where do you see that error20:40
Mister_Magisteri see errors that mal mentioned20:41
ghosalmartinmal, Mister_Magister when running mic Warning: repo problem: nothing provides pulseaudio >= 11.1+git4 needed by pulseaudio-modules-droid-11.1.73-1.1.4.armv7hl,20:41
Mister_Magistermaybe try fixing your repo firsty20:41
ghosalmartinthat shouldnt matter in this case should it? since ive commented out the packages that require droidmedia etce tc20:41
malghosalmartin: your OBS is failing20:41
ghosalmartinand done that20:41
Mister_Magisterthey wont rebuild if they contain error20:41
malghosalmartin: OBS has 4 failing packages20:42
ghosalmartinmal: there not really required, the packages that need them are commented out here
ghosalmartinbtw for android-8.0 is the standard audioflingerglue okay?20:42
malghosalmartin: why have you done that?20:43
ghosalmartinmal: because I just wanted a basic zip at the time20:44
malghosalmartin: are you using correct version for mic build?20:44
ghosalmartinalso this is the pulseaudio it pulls pulseaudio-11.1+git2-1.4.1.armv7hl.rpm20:44
ghosalmartinmal: hmm has that changed recently?20:44
malghosalmartin: 23:36 < mal> ghosalmartin: latest is now
ghosalmartinam being an idiot20:44
ghosalmartini forgot to update RELEASE=20:45
malMister_Magister: and about your help, you do help, now you were looking at the correct place for the reason for te error20:45
ghosalmartinmal: is htis the error that needs a correct version of mic?
malghosalmartin: as mentioned earlier today, update sdk20:49
ghosalmartinmal: will do, btw did you ever see the patch that I posted that fixed my OBS issue in end?20:50
malghosalmartin: so ssu re, zypper ref and zypper dup, note without sb220:50
malghosalmartin: what OBS issue?20:50
ghosalmartinmal: for a long time I kept getting 404s20:50
ghosalmartinit was some bug with the oauth system20:50
ghosalmartinlemme see if I can find the post again20:51
piggzmal: ok, so after getting the quirks to load....20:52
piggzheight = video plays, no crash, but colors wrong20:53
piggzheigth + width, width = broswer crash20:53
malpiggz: so now you need to figure out why the colors are wrong, check gst debug output20:54
piggzmal: well, it hands it off to gstdroid....20:57
malpiggz: are you sure it actually does any conversion20:58
ghosalmartinmal: this patch;bug=762825;filename=fix-urllib2-workaround.patch;msg=521:01
piggzmal: it definitely calls this function21:14
piggzand does not take the error path21:14
piggzabranson: ^21:14
malpiggz: so it's not using the gst-droid color conversion21:22
piggzmal: no ... not if I keep libI420colorconvert.so21:22
piggzif I remove it, it does use gst-droid conversion21:22
malboth have failing colors?21:23
piggzmal: yes, but in different ways21:23
piggzwith the lib = purble/green hints21:23
malpiggz: and you don't see "Unknown codec colour format"21:24
malwithout the lib21:24
piggzmal: well, i modified the if stmt to handle format 2121:24
piggzso no ;)21:24
malso maybe 21 needs some changes to the conversion21:24
piggzyes, i did some reseach21:25
piggzit could be (COLOR_FormatYUV420SemiPlanar)21:25
piggzdoesnt explain why the built-in lib fails!21:26
malghosalmartin: so osc had/has issues?21:26
ghosalmartinmal, yeah, it wouldnt work without that patch. was really annoying i figured it out once before but forgot21:27
ghosalmartinmal: btw do defconfigs inherit from other defconfigs?21:27
malghosalmartin: sometimes21:29
malghosalmartin: I think newer androids support split defconfigs21:29
ghosalmartinmal: hmm so I wonder if extra params are being added in somewhere. i best check .config21:30
malghosalmartin: the defconfigs are mentioned somewhere in device repos21:31
ghosalmartinmal: okays ta, ill dig into that21:31
piggzmal: not sure i can code a conversion of semi-planar to planar21:37
malpiggz: can't you adapt the packed semiplanar21:43
piggzmal: am i allowed to??? !!!!21:48
piggzmal: so, there are 2 planes afaict ... an now im off to bed, its a job for tomorrow!21:51

Generated by 2.17.1 by Marius Gedminas - find it at!