Tuesday, 2019-11-26

r0kk3rzThaodan: looks like we can just scp it from gitlab if you have an appropriate user01:20
T42Geekmar was added by: Geekmar03:38
Thaodanr0kk3rz: if you want I can create one later04:48
*** moralrec1rdings is now known as moralrecordings05:31
r0kk3rzThaodan: yeah i dont have much space on my server05:38
rinigusMorning! r0kk3rz , looks like that build went through. I was planning to have a pipeline with 6 builds in parallel (all variants of tama, one per job). Do you know whether looks are for the still of artifacts or is it some runtime limitation?05:56
rinigusRight now, I have uploaded build as a release file at https://github.com/sailfishos-sony-tama/main/releases . As it's aosp and below 2gb per file, it should be ok05:59
rinigusAs with the servers, it's frequently the bandwidth of meet connection that's a problem06:00
rinigusThaodan /\06:00
Thaodanmine  in a datacenter in Germany06:01
Thaodanit should be fine06:01
rinigusQuestion is in the bill usually. You probably have some number in the contract06:04
r0kk3rzrinigus: if its archiving ok then maybe they changed the rules since i was trying the x compact build06:07
rinigusr0kk3rz : will have to see the generated zip.06:08
r0kk3rzrinigus: something is definitely wrong there06:09
r0kk3rzsurely 39.5MB is not big enough06:09
r0kk3rza normal build is like 300 ish isnt it?06:09
Thaodanrinigus: You mean traffic?06:22
rinigusr0kk3rz: its usually more than 1GB. will have to look06:57
rinigusThaodan: yes06:58
T42<faenil> rinigus: thanks for the kernel configs, just as I was looking into it, yesterday 😁07:19
rinigusfaenil: sorry, should have warned. I am right now setting it up the builds for all the variants.07:22
rinigusconfigs are ready, as much as I can do without devices.07:22
rinigusfaenil: hopefully, the fun part will start when the builds are done and you could start testing. I suspect that touch device will give us a bit of a problem, as it may be different among devices07:23
r0kk3rzyou should be using udev for that anyway07:24
T42<faenil> rinigus: it's okay, it was helpful :) I am still going through the instructions to do the first build , plan to finish today...had some minor problems here and there07:24
rinigusfaenil: right now I have system/vendor for xz2 (dual) and xz2c(single) compiled, not uploaded to obs. as my src code contains gpt utils fixes, I'll better do all aosp compilations07:24
T42<faenil> What do you mean by all aosp compilations?07:25
T42<faenil> On obs you mean07:25
rinigusfaenil: system and vendor images for all 6 variants.07:26
T42<faenil> Oki07:26
rinigusno, that's done locally. packages are uploaded to obs07:26
T42<faenil> Ok07:26
T42<faenil> A build system that doesn't build? :D07:27
rinigusfaenil: it does build the rest07:27
rinigusfortunately, we are not going to rebuild those bits much. I hope07:28
T42<faenil> :D ok07:28
rinigusr0kk3rz and Thaodan : houston we have a problem. artifacts too large. so, how can we upload it to somewhere else?07:31
Thaodanrinigus: yes, please send me an e-mail with your public key and then I'll send you logindata later07:42
rinigusThaodan: thank you! I'll generate special one for gitlab07:43
deathmistmorning, how can I include https://github.com/sailfishos-oneplus5/nemo-qml-plugin-systemsettings in my builds by default? it stops my internal partitions being treated as external and after building it & adding the package to patterns, the build still uses the official version :/07:46
riniguslocal or obs? at obs, see tama project.07:47
deathmistjust local, I haven't figured out OBS stuff on the device side even tho the project should be fine (needs to be updated for 16.0 first tho)07:48
deathmistin fact I should probably do that today and attempt getting OTAs working (maybe test from 3.1 to 3.2, kernel subversion updates and how it affects built modules on device etc)07:51
rinigusdeathmist: didn't try that on local. but I presume you can build it as any mw07:52
r0kk3rzrinigus: if you send me the private key, i can put it into a shared secret thing between all the CI jobs07:54
rinigusThaodan: would that be OK? ^07:56
Thaodanso that every job can use the server?07:57
Thaodanthat would be ok07:58
rinigusI presume so07:58
Thaodanjust create a folder for each device07:58
rinigusr0kk3rz: your email, please. (don't know how to start PM via matrix on irc). send email via pm08:00
Thaodani assume their a three type's of builds just like on obs?08:01
rinigusThaodan: I know testing and devel. which one is the third one.08:06
rinigusnever happens :)08:07
rinigusI think ports always release from testing08:07
Thaodani think Sony's could be in release if the aosp support is good.08:10
Thaodanbut requires more manpower for more structured development08:11
rinigusThaodan: let's start with 2 and then see where do we get with it08:11
r0kk3rzmost devices never make it out of devel08:12
deathmistoh right I got stuck at paccopy before when I needed to basically clone devel to testing because of some username / permission issues, will have to take another look at that after I update all the packages at least08:17
deathmistabout naming of the project(s), http://www.merproject.org/logs/%23sailfishos-porters/%23sailfishos-porters.2019-11-19.log.html#t2019-11-19T07:20:55 still stands08:25
deathmistmal: r0kk3rz: ^ any opinions about that? everything currently lives at https://build.merproject.org/project/show/nemo:devel:hw:oneplus:cheeseburger08:27
malThaodan: about the [[]] vs [] in those script, the first line in those scripts defines we use bash so it there some issue using [[]] in that case08:57
ThaodanNot issues but it's not really needed imho.I tend to use the posix syntax most of  the time except when I need something like arrays or perl style regex marching09:00
ThaodanAlso there are features that are pretty useless like the function keyword09:03
malThaodan: ok, just thinking of that wildcard matching which would need [[]] or case09:26
Thaodani tend to use  case then09:41
T42龚莆婷 %lastname% was added by: 龚莆婷 %lastname%09:52
Thaodancan we invite shieldy into the telegram side like in the main group?10:06
T42<birdzhang> too many spam bots10:09
T42sheildy_bot was added by: birdzhang10:09
Thaodanthere are also bots who spam ssh servers there are far worse in terms of numbers10:15
piggzmal: aby idea how ive broken my target?     magic_load failed: File 5.14 supports only version 10 magic files. `/usr/share/misc/magic.mgc' is version 1411:54
piggzwhen buildng rpms11:54
malpiggz: did you update both sdk and target?12:31
abransonand tooling!12:44
Mister_Magistermal: u got time later today?12:45
malfor what?12:48
malabranson: I have never really understood how tooling is updated12:48
malabranson: the sdk-assistant update does usually nothing for that, not sure how that can be used together with changing version12:49
abransonmal: sdk-manage can do all of them12:49
Mister_Magistermal: debugging hybris1612:49
malabranson: I'm usually just lazy and remove the whole tooling and target and install again12:49
abransonpiggz: a good hint there ^ :)12:49
Mister_Magistermal: i think i finally solved all issues with asus stupid retarded bootloader so i wanna get my asus on hybris16 to gui, i got i think all services needed but gui it not starting and user session is freezed (you cannot ssh to nemo or su to nemo)12:57
Mister_Magistermal: so i was wondering if you have time to help me today12:59
malwe'll see in the evening, I will be doing new fp2 release later today but that is just writing release notes etc13:03
Mister_Magistermal: okay! i will write to you if i solve bootloader issue13:04
piggzmal: abranson: ok13:19
deathmistI still didn't get a response earlier today, so how should we deal with making my OBS projects for OP5/5T like https://build.merproject.org/project/show/nemo:devel:hw:oneplus:cheeseburger ? I've got a rough idea http://www.merproject.org/logs/%23sailfishos-porters/%23sailfishos-porters.2019-11-19.log.html#t2019-11-19T07:20:5514:16
deathmistmal: r0kk3rz: ^14:17
T42<nitin_3> @deathmist did you check?..were ssid showing twice in the list of wifi networks for you also?16:50
deathmist@nitin_3 yep I did, and no I only see each SSID once17:01
deathmistkimmoli: spiiroin: any clue why the gesture-daemon methods would seemingly trigger twice for everything? log only says "MCE signal power_button_trigger says event7" once yet the action is executed twice, here's my EV_MSC "hack" https://git.io/JeXv418:32
Mister_Magistermal: u there?18:32
deathmistmy fork of gesture-daemon https://git.io/JerMg and custom MCE config for gesture handling https://git.io/JeXv0 (if those matter)18:33
malMister_Magister: I was just about to go for a short walk, maybe describe the issue you have and provide logs and I'll check those after I get back in about half an hour or so18:38
Mister_Magistermal: go for a walk i magically got gui. will try to get logs and stuff18:38
Mister_Magisterbasically sfos boots but only telnet works su root or nemo freezes18:38
Mister_Magisteraand i don't have gui anymore after it rebooted itself18:38
malMister_Magister: make sure you enable persistent journal logging so you can see why it rebooted18:41
Mister_Magistermal: logcat is spammed18:41
malwell that looks bad18:42
malgrab dmesg or journal so we can see how the droid-hal-init works and if any services fail18:42
Mister_Magisterit's quite stable now18:43
kimmolideathmist: couldnt find anything wrong with a quick look18:47
maldeathmist: show what evdev_trace -t shows18:47
kimmoliif this hits only once?  https://github.com/sailfishos-oneplus5/gesture-daemon/blob/master/src/gesture-actions.cpp#L4218:51
kimmolisure you don't have two instances running?18:51
Mister_Magistermal: surfaceflinger fails with failed to create composer client but if i kill hwc i get bootanimation18:52
Mister_Magisterso test_hwc2 works now fine aswell18:54
Mister_Magistermal: in dmesg all i can see is IPC_RTR: msm_ipc_router_bind: pm-service Do not have permissions18:55
maljust show the whole log18:55
malMister_Magister: are you sure you didn't have the usual issue of eventd*.rc not appearing in out/18:56
Mister_Magistermal: i put eventd there18:56
Mister_Magisteras a confirmation all 3 servicemanagers are working18:57
Mister_Magisterhwc is working too… keymaster too…18:57
Mister_Magistermal: i can't… it's fully spammed 11 secs after boot19:00
Mister_Magistercan't get dmesg ealier19:00
spiiroindeathmist: I can't spot any obvious faults either19:06
T42<faenil> rinigus: --version build fails with unsatisfied BuildRequires, all the droid-config-*.19:09
T42<faenil> Any reason why the droid-configs build has --do-not-install?19:09
T42<faenil> Also all the droid-hal-* are missing19:09
T42<faenil> I wonder if the instructions are missing some install commands, or I am missing something :)19:15
malMister_Magister: that is why I asked for persitent journal log, the kernel messages and droid-hal-init stuff goes there also19:20
Mister_Magistermal: i got bit further with android devs already :)19:20
rinigusfaenil: I had akari in few instructions specified, corrected with the last commit in hadk19:24
rinigusfaenil: that droid-config will have to be fixed.19:25
deathmistok uuuh now the gestures just started working properly by themselves automagically? I didn't even change anything since it didn't work just moments ago ._. will do more testing... "evdev_trace -t" just gave me the MSC_GESTURE and SYN_REPORT events as expected from the changed source19:25
rinigusI am right now building hal for xz2c single sim, should be ready soon19:25
rinigusandroid is ready for it, not uploaded yet19:25
rinigus10-15 mins19:25
rinigusfaenil: are you trying to build it locally or get from obs?19:26
Mister_Magistermal: okay i know how to fix it now but got problem19:30
Mister_Magisterhttps://github.com/mer-hybris/android_system_core/blob/hybris-15.1/init/README.md#options capabalities is not workign19:30
T42<faenil> rinigus: all locally. I noticed the akari commands yesterday and had a PR ready but then I saw you already committed today :D19:34
T42<faenil> The android build is done already19:34
rinigusfaenil: then that's good. its great if you get through as well - the more checks the better19:35
T42<faenil> Yeah I don't like fast food ;)19:35
rinigusthe first droid-configs are built with install.19:36
T42<faenil> Just noticed that the boot img failed to build, there's a conflict between busybox-symlinks-cpio and the cpio BuildRequire'd by the boot img spec19:36
T42<faenil> I'll have a look later, dinner time o/19:36
T42<faenil> Yeah noticed that :/19:36
rinigussee https://github.com/sailfishos-sony-tama/main/blob/master/hadk-sony-xz2.md#build-packages19:36
T42<faenil> Yeah I ran those commands again just to be sure...19:37
Mister_Magisterrinigus: do you have something to drink? we got something to celebrate19:37
T42<faenil> And then as I building boot img again I noticed the cpio conflict19:37
T42<faenil> Bbl19:37
rinigusMister_Magister: not tonight, unfortunately. but I am listening...19:38
Mister_Magistersec lemme write tweet19:38
Mister_Magisterrinigus: mal: https://twitter.com/Mister1Magister/status/119941233435314585619:39
rinigusMister_Magister: congratulations!19:42
T42<faenil> Congrats!19:42
Mister_Magisterrinigus: thanks! i went through hell to get it19:42
Mister_Magisterwifi is working ootb xd19:44
Mister_Magisternow time to write notch support19:45
rinigusI can see that. even started chewing on one side of the screen and now you will miss the time on status bar19:45
T42<nitin_3> wow..congrats...it also has sdm845 right?19:45
Mister_Magisterrinigus: yeah i will write support19:45
Mister_Magister@nitin_3 hell yeah it is19:45
Mister_Magisteri wanted to be first but i ain't first anymore19:45
Mister_Magisteronly because asus19:45
T42<nitin_3> why..what happend19:46
Mister_Magister@nitin_3 adding android and spl version to bootimg with buildv property was causing my bootloader to freeze and since i didn't have working bootctl i ran out of tries and bricked my device so i had to remove back glass in order to short pins to get it to edl and then flash bootimage to get bootloader working again and after it all i learned that lineage devs added buildv property to kernel cmdline ONLY to fix Error value on19:49
Mister_Magisterbootloader screen and that buildv property was causing bootloader to freeze!19:49
Mister_Magisterah i forgot to add i have to add android and spl version just because asus devs were too lazy to remove this requirement (as google removed it)19:50
Mister_Magister so annoying19:50
Mister_Magisterwritting about it is short but actually fixing it…19:51
T42<nitin_3> do you think it is intentional?19:57
T42<nitin_3> ^was19:57
Mister_Magisternah other guys on other vendors had same lame lazy vendors19:58
Mister_Magisterbut asus bootloader is snowflake19:58
malMister_Magister: nice that you got UI working now, do logs look better now?20:18
Mister_Magisteryeah but i will explain it to you how i fixed it20:18
Mister_Magistermal: i need to add capabilities NET_BIND_SERVICE to pm-service but it didn't work so i just disabled checking for permissions in kernel. so i just need to fix capabilities NET_BIND_SERVICE and it will work fine20:19
rinigusfaenil: I corrected the command in hadk for config build, in root fs section20:27
rinigusfaenil: mountpoints are the same for akari and apollo - https://github.com/mer-hybris/hybris-boot/pull/17920:29
T42<faenil> rinigus: yeah I already defaulted to using the same mountpoints :) haven't even unlocked my device yet, so haven't checked on aosp20:35
rinigusfaenil: just uploaded hal and system/vendor for apollo single sim. as soon as system/vendor rpms are prepared by obs, droid-hal-version-h8314 should be resolvable (in devel https://build.merproject.org/project/monitor/nemo:devel:hw:sony:tama)20:36
rinigusfaenil: do you have single or dual sim?20:37
T42<faenil> Cool. I need to get it done locally though :P or I won't enjoy/learn20:37
T42<faenil> Single20:37
T42<faenil> I don't think the configs spec was the problem, it was building all packages already :/20:38
T42<faenil> I think the cpio conflict is the problem atm, I'll fix that first20:38
rinigusfaenil: that special feeling when you baked it yourself...20:38
T42<faenil> :D more about learning, but yeah20:39
rinigusfaenil: it maybe was building too many. try with that command20:39
T42<faenil> No no, it built all of them.20:40
T42<faenil> I tried with the new cmd, no difference20:40
T42<faenil> I think it's just that it doesn't find the boot-img package (which is not building due to cpio) and the missing droid-config is just a false positive due to the missing  boot img pkg20:41
rinigusfaenil: not sure I understand where the issue is. pastebin some logs ?20:42
rinigusbut config was built before boot-image. did you remember to run sb2 command before boot image?20:42
T42<faenil> The failure is not in config, it's in boot-img and version20:43
T42<faenil> I'm writing from the phone, it's faster if I just fix it and then write here :)20:44
rinigusfaenil: sure! good luck20:46
T42<faenil> Cheers20:48
malMister_Magister: quite strange issue20:48
Mister_Magisterthis whole device is strange20:49
T42<faenil> Rinigus: boot-img build fixed by manually installing cpio in the sb2 target and manually resolving the installation conflict (I chose to remove busybox-symlinks-cpio, replacing it with cpio)20:50
T42<faenil> Back to version now, still complaining about missing providers for all the droid-hal-* and droid-config-*20:52
vknechtin case different tama models have different partitions layout, overriding flash-partition to use MER_HA_DEVICE could be useful, like here for loire https://github.com/vknecht/droid-config-sony-loire/commit/f8ac0dfc5deafd2dad8ff24c4a2edd91b0a2c4d121:04
vknechtalso restore /dev/touchscreen, to support different /dev/input/eventX depending on model: https://github.com/vknecht/droid-config-sony-loire/commit/e76ddd126ccf807ff4bc3c534b736dcb573b62d721:05
T42<faenil> rinigus: mmm I see droid-configs.inc provides droid-config-%{rpm_device} but also conflicts with bluez5. Both are build deps of the -version pkg though21:07
T42<faenil> is that expected?21:07
rinigusvknecht: let's see. I hope we don't need it21:07
rinigusfaenil: strange, you should have those in local repo21:07
T42<faenil> I am just reading the .spec21:07
rinigusvknecht: touchscreen I may need. xz2c maybe different21:09
T42<faenil> rinigus: built —configs for the 3rd time...after that —version built successfully... I am confused. —configs succeeded in the past already21:11
T42<faenil> unless the script is not reporting a silent failure21:11
vknechtrestoring /dev/touchscreen was kind of easiest/laziest, maybe cleaner to have per-device  sparse/var/lib/environment/compositor/droid-hal-device-....conf21:12
rinigusfaenil: good that it did.21:12
T42<faenil> yeah, I guess I'll just assume I did something wrong twice21:12
rinigusvknecht: I was told that xz2c has different batches with different touchscreens21:13
rinigusdon't know if its different event dev21:13
T42<faenil> and now again, can't find jolla-configuration-h8314 ... argh :)21:14
vknechtat least for X vs. Xcompact, it was a different one ; also,maybe tama isn't only "clearpad" driver, at least both suzu and kugo are (fwik)21:15
rinigusfaenil: that jolla-configuration- is built using do-not-install variant of config. do you have it as a package in local repo?21:17
T42<faenil> tried building the rootfs twice, once failed with jolla-conf-$device not found, then with droid-config-$device-bluez5 not found (needed by jolla-conf-$device).21:18
T42<faenil> oh well21:18
T42<faenil> rinigus: second time it was missing bluez5 droid config21:18
rinigusfaenil: I am off to sleep. you could also cheat by pulling KS from http://repo.merproject.org/obs/nemo:/devel:/hw:/sony:/tama/sailfish_latest_armv7hl/armv7hl/droid-config-h8314-ssu-kickstarts-0.2.0+master.20191124153518.1.g61a77f1-1.2.1.jolla.armv7hl.rpm21:19
T42<faenil> so it seems like during the second run it could find jolla-conf (how? it couldn't find it on the first run...) but it failed to install it21:19
T42<faenil> rinigus: no worries, I'll play with it and get it fixed :) have a good night!21:19
T42<faenil> rinigus: if you're still around, just one thing, what is supposed to build droid-config-$device-bluez5? I can't find any provider for it in .specs21:23
T42<faenil> found, nvm21:26
TheKitcan I ask for planet-cosmocom on OBS?21:45
r0kk3rzTheKit: thats the codename?22:17
r0kk3rzor planet:cosmocom ?22:17
TheKitr0kk3rz, thank you22:29

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!