Friday, 2019-01-11

*** ChanServ sets mode: +v T402:01
T4<xreactx> Is it possible to make Sailfish OS like an Android Generic System Image? So it can be installed on any treble device?02:03
*** ChanServ sets mode: +v T408:06
*** greguu1 is now known as greguu08:12
Thaodanis ubu-chroot no longer insideb the chroot?08:41
piggzThaodan: install android-tools-hadk09:02
Thaodanpiggz: thanks, I try zypper doesn't respond09:48
T4<adampigg> Thaodan, hmm,.ought to ...not at the computer atm.....does it timeout or anything?10:00
r0kk3rz@xreactx i dont think so, at minimum we require kernel changes, which isnt covered under the GSI10:05
Thaodannope just doesn't react even on simple querys10:05
Thaodanrpm worjs fine thow10:05
piggzperhaps strace?10:55
Thaodantried but doesn't show much:11:01
Thaodanany idea?11:01
piggznot really tbh11:05
piggzmal: roughly when is 3.0.1 out for general/ports/obs?11:14
malgood question, it's EA now so probably not too long, usually target comes after real release, sometimes we get EA11:17
piggzmal: interesting, this writeup says how to get raw gyro data from my watch ... i dont thing gadget bride implements this, but some android apps must something similar for creating actions from watch movements
piggzwhat im actually looking for, is how to interpret the sleep data11:33
Thaodanpiggz: no idea why its that hell slow its working but slow no idea how this can happen on a clean chroot11:43
piggzThaodan: you are still on a 56k modem?11:43
Thaodannope its only zypper11:44
Thaodanbut 50mbits is still like 56k compared to the rest of the world11:44
Thaodan[adaptation0|plugin:/ssu?repo=adaptation0] Valid metadata not found at specified URL wtf?11:46
r0kk3rzthat one is fine11:47
malThaodan: ssu lr, that is the url for that?11:53
Thaodanthat was the zypper url11:54
Thaodanfailed on too11:54
malThaodan: are you running zypper in sdk or target?12:00
Thaodanin sdk12:00
Thaodanzypper log12:03
Thaodando I need to do something to use zypper after extracting the archive?12:04
Thaodanbecause of  the zypper : [adaptation0|plugin:/ssu?repo=adaptation0] Valid metadata not found at specified URL message12:05
malThaodan: is that a fresh install of sdk?12:10
malI assume you mean platform sdk, not the app sdk12:11
Thaodaneh yes and yes it is12:11
malThaodan: for me it worked fine, both and
malsame url12:17
malThaodan: so zypper ref fails also?12:18
piggzThaodan: do the urls work ok from your PC?12:30
Thaodanexcept the 403 forbidden yes12:32
malpiggz: those never work from computer12:32
malpiggz: using browser that is12:32
piggzbecause of the auth?12:33
Thaodanmal:  thats what I did:
Thaodanstuck at the last command now12:34
malThaodan: what kind of internet connection do you use?12:44
Thaodan50mbits dsl12:44
malI don't quite understand some of the lines, what runs that zypper command?12:46
T4<It_sMike> mal any idea about this?12:48
mal@It_sMike what are you trying to do?12:49
mal"unknown filesystem type 'squashfs'" means you don't have needed kernel patches for anbox12:50
T4<It_sMike> Oh ....12:50
T4<It_sMike> which patches should I add to kernel ?12:51
T4<It_sMike> and also this problem : , is it returns to kernel? Because I don't think so...12:51
Thaodanmal: a function that calls the mer-sdk-chroot script12:57
malThaodan: something you did or builtin?13:00
mal@It_sMike anbox is not in a good shape now13:01
T4<It_sMike> Yeah13:01
r0kk3rzput the patches in though, its nice to have some devices people can use for anbox development13:28
T4<gischpelino> piggz: sleep detection you have this source? …
piggz@gischpelino i think ive read that before ..... i guess i need some custom processing over and above just doing an SQL query like everything else so far!13:34
T4<gischpelino> I tried Gadgetbridge on lineageos last December, it still has problems to detect deep sleep. I read the data from the device are not well dokumented and still not hacked completely.13:38
T4<adampigg> yes, i store the raw data from the device in a database (you can view it yourself) and can pull out certain things like heartrate, steps etc, but there are a load of other activity types and theyre not well documented13:39
piggzmal: how can i ship a change to a ro property?13:48
piggzactually, i might have it13:51
piggzi had a conflicting change in /system/build.prop13:51
*** ChanServ sets mode: +v T414:09
malpiggz: which one of the property files is preferred, device repo can add overrides14:11
dcalistemal: as you may have seen already, I've added some more information about the crash in Mer SDL2 MR. I think the issue is a design problem in SDL and I'm proposing a patch to solve it. When you have time, may you test it further and tell me if it actually fix the crash? In my simple tests it's the case, but it's better when looked by others!14:25
T4<adampigg> Mal, the property is
maldcaliste: yes, I saw the comments but I haven't had time to thing about that yet, I'll have a look tonight or this weekend14:33
maldcaliste: maybe we should join sdl IRC channel and ask their opinion? not sure if any other system uses rotation with sdl wayland?14:36
maldcaliste: I'm not 100% sure if libhybris wayland backend is following the standard very well14:39
dcalistemal: yes take your time for testing there is no hurry. About libhybris, you may be right, but read at the specifications, I understand them as it is now: output events are emitted at every change and not only on discovery.14:43
dcalisteBesides, you're right, I should consult SDL developers also. I'll do later after you will have the opportunity to test and see if it's indeed curing the issue.14:44
maldcaliste: ok, thanks for helping debugging the issue, I have too many projects so I haven't had time to look into wayland specifications14:45
maldcaliste: but considering that we have no problems with Qt crashing the issue could indeed be in SDL14:48
maldcaliste: wondering if your change to display_handle_mode prevent updating display size?15:01
dcalistemal: yes I'm wondering the same. But I guess not because the current code cannot be used to signal device rotation. Maybe it is broken in SDL 2.0.8 but it is then broken also without the patch (which cannot be tested of course).15:08
dcalisteI mean, it is the transform argument from the geometry signal that is giving the orientation information, but this argument is not used.15:10
dcalisteWhat I'm guessing is that this code is only used to get the wl_output structures stored as driverdata and then somewhere else there is some code to listen to changes to these structures.15:11
maldcaliste: master sdl has orientation handling in android backend, maybe similar could be done in wayland backend15:12
malSDL_SendDisplayEvent(display, SDL_DISPLAYEVENT_ORIENTATION, orientation);15:13
dcalisteYeh, I will check from looking at the sources that this patch has no link with the device rotation detection.15:17
malyes, but as a future idea15:23
T4<dashinfantry> @piggz [ …], Need  a ticwatch one15:50
piggz@dashinfantry send me a ticwatch :D17:00
piggztho, seriously, need more devs, implement AbstractDevice to create new device types in Amazfish17:01
Benclark006evening all17:17
Benclark006my device is bootlooping, is there a fix for it?17:17
Benclark006i'll grab some logs in a sec17:17
T4<It_sMike> Evening, did you disabled SELinux?17:18
piggzBenclark006: love the vagueness of your problem! :D17:18
Benclark006@It_sMike i have17:18
T4<It_sMike> :) so .. Send any logs to see what's wrong17:19
Benclark006piggz: yeah, i have no idea what's going on so i can't be to specific17:19
Benclark006Well, here's the system journal:
Benclark006and dmesg:
piggzBenclark006: the answer may be in your kernel cmdline17:26
Benclark006piggz: how do i access that17:46
Benclark006sorry for the dumb question17:46
piggzBenclark006: you pasted it!18:01
malpiggz: that has selinux=0, assuming the kernel parameter is set correctly18:14
piggzmal: look at the boot reason though18:14
piggzit was a watchdog reboot18:14
malpiggz: is that what is seen in selinux issue?18:14
malBenclark006: could you show your kernel defconfig again18:15
piggzdoesnt that mean that mce didnt set the watchdog in a timely manor, and the kernel reboot itself ?18:15
malnot sure18:16
piggzthat it what I had happen on latte18:16
malpiggz: how did you fix it?18:17
piggzthere were 2 fixes....18:17
piggz1. the one i currenty use is
piggzbtw s/mce/dsme18:19
piggzdisable DSME watchdog writing becuase it doesnt work when suspended .... the android WD daemon does work while suspended18:19
piggz2. the one @eugenio_g7 uses was to patch our kernel so tha when the device suspends, it doesnt need the watchdog tick18:20
malpiggz: afaik mako device didn't need any hacks before18:20
piggzmal: ok, i might be wrong, but the commandline certainly suggests something like that is happening18:26
piggzis droid-hal-init starting? are there any watchdog proceses in /init* ?18:27
Benclark006hi sorry i went out for a bit18:27
piggzBenclark006: what is the time taken for a reboot to occur?18:28
Benclark006i'll take a look18:28
malpiggz: if it was selinux it should be quite fast18:31
malBenclark006: maybe use dmesg -w and journalctl -f to follow logs as long as you can18:31
Benclark006it takes 1:32 until it reboot18:31
piggzBenclark006: also, post link to config, and, look in /sys/fs/pstore/*18:32
malpiggz: too long reboot time for selinux, maybe modem?18:32
Benclark006that sounds plausible18:35
Benclark006here's dmesg -w
piggzthat shows ~46 seonds, not 90 ;)18:39
Benclark006i timed it18:39
Benclark006maybe i did it wrong 🤔18:39
malpiggz: as you can see last line it didn't follow the log18:39
piggzBenclark006: do you have pstore?18:40
Benclark006journalctl -f:
Benclark006i don't think so18:41
Benclark006how do i check?18:41
piggzmal: what causes a modem to reset the device/18:41
malpiggz: I think the issue could be something else, look at the android services being killed18:41
r0kk3rzmodem decides to draw 10A?18:42
piggzmal: should it load .rc files from /usr/libexec and /system?18:43
malpiggz: that is 15.1 which uses .rc files differently18:46
malBenclark006: waiting for journalctl -f18:46
Benclark006ah, yes18:46
Benclark006i did send it18:46
Benclark006here you go
malnext time journalctl -f -n1000018:47
piggzalso, when logged in, humour me and do grep watchdog /*18:48
malBenclark006: you didn't show your defconfig again18:48
Benclark006oh heck18:48
Benclark006give me a sec18:48
malBenclark006: you really should update your github repos, but in config repo this time only add file to sparse (git add sparse/*)18:49
Benclark006good idea18:49
Benclark006^that's the defconfig18:50
Benclark006and the pstore directory doesn't exist18:51
malBenclark006: about that kernel config, you have set CONFIG_AUDIT=n which is not what kernel config checker said, it said that you use CONFIG_SECURITY_SELINUX_BOOTPARAM=y then you keep CONFIG_AUDIT=y18:52
Benclark006ok, change that18:53
Benclark006anything else?18:53
Benclark006ok, i'll build it and let you know19:01
malBenclark006: you only need to run make hybris-boot and the use fastboot to flash the hybris-boot.img to your device19:06
Benclark006where's that image?19:07
piggzit tells u in the output19:08
malsomewhere under out/19:09
piggzflash hybris-boot.img, not boot.img19:11
Benclark006ok neat i've done that19:13
piggzlooks like there are some worthwhile account/calendar improvements in 3.0.1, looking forward to it19:13
Benclark006doesn't look as if it's bootlooping19:15
Benclark006it hasn't rebooted yet19:16
Benclark006but telnet isn't working19:17
Benclark006could it still be running?19:17
malcheck host dmesg19:18
Benclark006unable to enumerate USB device19:19
malprobably reboots, or the android side usb handling breaks things19:20
malBenclark006: add back CONFIG_AUDIT=n19:26
Benclark006should i keep CONFIG_SECURITY_SELINUX_BOOTPARAM=y?19:27
malkeep other things they are now19:29
*** ChanServ sets mode: +v T420:10
Nokius:-/ what was the fix for geoglue in the cancro target o.O20:22
Nokiusah the token section20:22
malNokius: package name20:23
malNokius: which should match one of spec file names if there is more than one spec files20:24
T4<It_sMike> A 's' was missing :D , also downgrade to an old version20:24
Nokiusmal: ah okay thx20:24
malNokius: do you have some build that is broken?20:25
Nokiuscancro was20:25
Nokiusmal: may find5 is I sould give it some love I know20:25
malNokius: looks like find5 OBS project hasn't build updated in a long time20:27
Nokiusmal: I know I'm so so sorry20:30
Benclark006so i've flashed the new image20:34
Benclark006it and i think it's working ok20:34
malBenclark006: working how?20:42
Benclark006i mean it doesn't bootloop anymore20:42
malBenclark006: so get logs20:42
Benclark006ok here's journal:
Benclark006and dmesg
malBenclark006: how long after boot did you take that journal?20:51
Benclark006about 10 minutes20:51
Benclark006should i do it a tad sooner?20:52
malno, just very few messages which is odd20:52
malBenclark006: did you remove init_user0 again?20:52
malfrom /init.rc20:52
malif I knew you are building new image I would have said how to remove it from the build20:53
Benclark006i don't believe so20:53
malyou need to remove it20:53
Nokiusarg I missed up sailfishos_2.1.3.7 in testing fu20:54
Benclark006all done20:54
malBenclark006: wait20:54
malBenclark006: do you have init.mako.rc in /20:54
Benclark006i'll check20:54
malor somewhere20:54
Benclark006i do20:56
malcomment out line "import init.mako.usb.rc" near the beginning of the file20:56
malBenclark006: also run systemctl mask ofono20:57
malNokius: why is the old building in cancro testing?20:58
Nokiusmal: it was not disabled :(20:59
piggzNokius: junior error :D20:59
piggzwe;ve all been there!20:59
malNokius: I disabled it20:59
Nokiusmal: will it stop do build now?20:59
NokiusI was wating and praying osc can revert to the old source :-/21:00
Nokiusand re run all :-/21:00
T4<It_sMike> May I help?21:00
malNokius: some of the packages were already rebuilt21:00
Nokiusyes with slapping my **21:01
malNokius: but maybe just add target21:01
Nokiusmal: I noticed21:01
Nokiusmal: yeah21:01
malunless you really want to make that old target work again?21:01
T4<It_sMike> :( sorry21:02
Nokiusmal: I added
Nokius@Mike mark as broken on obs and OTA should be avoided21:04
Nokiuspiggz: it's my secound time promoting :D21:06
Nokiushope it's still okay to do this rocky one21:06
T4<It_sMike> so everything's ok by now?21:06
Nokiusjust the repo for is messed up21:07
Nokiusobs is still building21:08
T4<It_sMike> :D guess there would be no problem, but people who wants to stay at
T4<It_sMike> Did you update droid-configs on obs? Just asking for sure!21:09
T4<It_sMike> testing repos I mean21:09
NokiusI promoted all pkg's to testing21:13
T4<It_sMike> Fine !21:14
T4<It_sMike> Let me know if anything's failed21:14
Benclark006ok so now it's bootlooping D:21:15
Nokiusit should if so my weekend is done21:15
Benclark006after 50 seconds the backlight turns off and the after another 50 seconds it reboots21:15
T4<It_sMike> :)21:15
Benclark006journalctl -f
malBenclark006: what did I say about getting journal log21:17
Benclark006oh yeah21:18
Benclark006hold on21:18
Benclark006well here's dmesg while i wait for my phone to reboot again
malbut anyway the problem seems to be android side21:19
malcan't remember what the fix was for the error that is seen21:19
malI remember others having the same issue21:19
Benclark006here's the extra logs
Nokiusfor the record cancro testing is disabled =)21:25
malBenclark006: not quite sure why it fails, did you keep a copy of the kernel with CONFIG_AUDIT=y?21:34
Nokiusr0kk3rz: thanks for the ci :)21:34
Benclark006mal: i think i've still got the image21:41
Nokiusr0kk3rz: failed for me but I assume why21:42
Nokiusr0kk3rz: how do you load the RELEASE and EXTRA_NAME?21:43
r0kk3rzremove that file from your sparse i guess21:43
malBenclark006: keep those different kernel versions so it's easier to test21:44
T4<It_sMike> Nokius the problem is droid configs21:46
Nokiusr0kk3rz: tell me :D21:46
T4<It_sMike> It's not updated with the one that I send for you21:46
r0kk3rzextra name is in the env file21:46
r0kk3rzrelease is set globally for all ci jobs21:46
Nokiusr0kk3rz: I don't get it, no extra is over writen too21:47
Nokiusr0kk3rz: where is it set in the pipeline?21:47
Nokiusr0kk3rz: okay I will give you sweeties in BRU and let you go when I know how it works :P21:48
r0kk3rzlook at the ci yaml file21:49
malNokius: you should tag the config repo21:49
malNokius: it's not latest in testing21:49
r0kk3rzyou got the before script line where it sources the .env21:49
r0kk3rzthen it runs the build shell script, and thats it21:49
malNokius: neither is it ok in devel21:49
Nokiusmal: merci21:50
malI enabled testing for thatr21:50
Nokiusmal: thx21:50
Nokiusr0kk3rz: I got this :P but How is the RELEASE getting in?21:51
Nokiusvs .Info[01/11 21:33:35] : Substitute macro variable @EXTRA_NAME@ with -Beta121:52
Nokiusfrom the ci job21:52
r0kk3rzyeah thats weird, it shouldnt be that21:53
r0kk3rzrelease is defined globally, so in an area only i can see21:53
r0kk3rzso i can set 1 field, and start clicking jobs, and boom updated builds for everyone21:54
Nokiusr0kk3rz: tell me more in BRU thanks in advance21:57
Nokiusmal: should be fixed for devel by now21:57
malNokius: both are building now21:57
r0kk3rzmany beers shall be had21:57
malr0kk3rz: heh21:57
r0kk3rzyou can bring your scorpion and we can be buddies21:58
Nokiusr0kk3rz: sure :D21:58
r0kk3rzmal: you should try and make it to fosdem, its a good time21:58
Benclark006i'm probably going to try again with 14.122:01
Nokiusmal: yes mal so it!22:01
Nokiusfor the record cancro testing is disabled again22:02
malr0kk3rz: when is that?22:05
r0kk3rzfirst weekend in feb22:06
r0kk3rz2nd, 3rd, but beer event is on the friday night22:07
Nokiusmal: take kernlyng with you :P22:07
maldepends on if I manage to fix my sleep issues or not22:15
Nokiusr0kk3rz: woop woop22:15
NokiusUploading artifacts...22:15
r0kk3rzdun dun dun22:16
r0kk3rzhmm, why does it have the wrong extra name, the others worked22:17
malit's good that some older ports are revived22:17
malBenclark006: not sure if 14.1 is easier or not, 15.1 is still a bit problematic sometimes22:18
malBenclark006: if you have enough disk space I suggest keeping the 15.1 sources22:18
Benclark006yeah i will22:19
malBenclark006: I think if you just copy the current android source tree locally and do repo init with 14.1 branch it should be able to minimize any downloads22:20
r0kk3rzit also says scorpion, ooops22:20
Nokiusmal: It's some much fun to work on it22:22
r0kk3rzhmm, were you retrying an old job?22:23
Nokiusr0kk3rz: si22:26
r0kk3rzah ok, it wasnt pulling in the latest stuff in the git22:27
Nokiusr0kk3rz: ah22:28
r0kk3rzthe way i do it is pipelines > run pipeline22:29
r0kk3rzbut now adding a tag will do it too, which once the job is complete it should list the artifact.zip22:30
r0kk3rzpiggz: apparently russians are using the gitlab CI as a mido bug tracker ;)22:32
piggzbut, i thought there were no bugs?!?!!22:32
r0kk3rzthis is an amazing one -
malprobably not many ports have a proper bug tracker like fp2 does
malr0kk3rz: wondering if some app pulls in old gst or something?22:33
piggzmy XDA page says to use droid-config-mido repo on githb22:34
malas mentioned earlier on this channel it really seems that the SDL crash issue which has been preventing wayland and SDL update is really a bug in SDL22:35
malit's now very clear where is happens and why22:35
piggzmal: so, you can ship a fixed sdl?22:36
r0kk3rzsounds like you've figured it out at least22:36
malmostly it was dcaliste based on initial information I gave22:37
r0kk3rzoh cool22:37
malI knew the line where it crashed but dcaliste figured out why that happens22:37
malit seems the people who did the wayland backend for SDL didn't have a device that can rotate it's display22:38
malI'm going to try to talk to SDL devels tomorrow to see what they think22:41
Nokiusr0kk3rz: does this make any sense ?
Nokiuscould be use full for images to share23:01
r0kk3rzwe were just setting up another branch23:05
Nokiusr0kk3rz: :D23:05
malNokius: is that enough? testing uses versioned targets but devel doesn't23:05
r0kk3rzbut in general i was only going to do devel stuff, testing images should be tested and released somewhere more stable23:06
malNokius: maybe you could have two .ks files? one for devel and one for testing23:06
r0kk3rzlike devaamo or something23:06
Nokiusmal: my bad, to much out of the biz23:06
Nokiusr0kk3rz: :D I wanna use someones pc to bild it :P23:07
r0kk3rzyeah thats fine, if you want to build it on there then you can23:07
r0kk3rzhowever you want really23:07
Nokiusr0kk3rz: getting my build systems up again will take ages :-/23:12
r0kk3rzinstalling the mersdk isnt that hard23:12
Nokiusanyway time for a nap oncall week is soon I should catch sleep gn823:12
malonly platform sdk should be enough for image builds23:15
Nokius;) yeap23:18

Generated by 2.17.1 by Marius Gedminas - find it at!