*** rinigus <rinigus!~rinigus@tabasalu-gw.radionet.ee> has joined #sfdroid | 07:09 | |
nh1402[m] | Why don't we have it wait for uinput during the boot process in system server for example, | 09:26 |
---|---|---|
krnlyng | dr_gogeta86: do you still get the boot message error currently? can you test something for me? | 11:10 |
rinigus | krnlyng: I tried to enable the sfdroid using your latest packages, but it seems to be stuck on CM boot logo when I press sfdroid. did you have any success on n4? | 11:12 |
krnlyng | rinigus: it should "just" work on n4 actually | 11:13 |
krnlyng | rinigus: can you check logcat for errors? | 11:13 |
krnlyng | nh1402[m]: we do exactly that | 11:13 |
rinigus | krnlyng: :( . I was just going to ask you how to debug the beast. OK, I'll reboot and try again. btw, which CM base did you use? | 11:14 |
krnlyng | rinigus: the one ballock suggested | 11:14 |
krnlyng | rinigus: YOG4PA iirc | 11:14 |
krnlyng | let me check | 11:14 |
krnlyng | yes | 11:15 |
krnlyng | piggz: ping | 11:16 |
rinigus | krnlyng: my image is based on YOG7D, as its default for hybris-12.1 [hence separate CM zip]. | 11:16 |
krnlyng | hmm | 11:16 |
krnlyng | that might cause problems | 11:16 |
rinigus | krnlyng: but probably we would use YOG4PA for release. | 11:16 |
krnlyng | ok | 11:16 |
krnlyng | is there a reason for using the one over the other? | 11:16 |
rinigus | krnlyng: so, if I want to test sfdroid, I might have to compile myself with the correct base, right? | 11:17 |
krnlyng | rinigus: either you use YOG4PA as base cm, or you compile sfdroid for YOG7D | 11:17 |
rinigus | krnlyng: /cross-posting/ advantage of yog4pa is that it was used for mako release | 11:18 |
rinigus | krnlyng: ^CM release, I mean | 11:18 |
krnlyng | ah ok | 11:18 |
rinigus | krnlyng: advantage of YOG7D is that its a base for hybris 12.1 | 11:18 |
rinigus | krnlyng: but then I have to compile CM as well | 11:18 |
krnlyng | rinigus: can you check whether logcat prints anything useful? the maybe there is just a small incompatibility which can be fixed ... | 11:19 |
rinigus | krnlyng: let me see | 11:19 |
krnlyng | rinigus: it probably rotates around like hell , capture it for like a minute and then look for the loop | 11:22 |
*** rinigus-m <rinigus-m!~sailfish@tabasalu-gw.radionet.ee> has joined #sfdroid | 11:40 | |
rinigus-m | krnlyng: i | 11:41 |
krnlyng | rinigus: u | 11:42 |
rinigus | krnlyng: I'll check the log a bit later - sorry, have to go :) [u] | 11:42 |
krnlyng | ok :) | 11:42 |
rinigus | krnlyng: I tried to cleanup the log, but since I don't know exactly what to expect, its rather long. here it is https://ptpb.pw/aVBk/#L-393 | 14:01 |
krnlyng | rinigus: ok another case of the showbootmessage error | 14:03 |
krnlyng | i don't understand why i don't get that | 14:03 |
krnlyng | rinigus: can you test this for me: | 14:03 |
krnlyng | https://bpaste.net/show/a6d8945b08e9 | 14:04 |
krnlyng | rinigus: this is init.sfdroid.rc^^ | 14:04 |
rinigus | krnlyng: just few differences with ballock's images: in addition to the CM base difference, I use a somewhat different kernel config (tried to add more things from new 12.1 CM kernel config). | 14:05 |
rinigus | krnlyng: sure, testing ... | 14:05 |
krnlyng | more android kernel configs cannot hurt :) | 14:07 |
rinigus | exactly! kernel seems also have got some updates by CM, so, while looking for playback bug, I baked a new config | 14:09 |
*** rinigus-m <rinigus-m!~sailfish@tabasalu-gw.radionet.ee> has quit IRC (Read error: Connection reset by peer) | 14:10 | |
rinigus | krnlyng: still stuck on boot logo (when I press sfdroid multiwin) | 14:15 |
krnlyng | hm ok | 14:15 |
krnlyng | same error in logcat? | 14:15 |
rinigus | krnlyng: how long I am expected to wait after boot? | 14:15 |
rinigus | which error do you look at? | 14:16 |
krnlyng | showBootMessage | 14:16 |
krnlyng | null object | 14:16 |
krnlyng | rinigus: it might take a while on first boot | 14:16 |
krnlyng | rinigus: but if it's not starting to optimize apps soon there is likely something wrong | 14:16 |
rinigus | krnlyng: let's not forget that this is a wrong base as well. | 14:17 |
krnlyng | rinigus: yes but this error is something that many people have | 14:17 |
krnlyng | rinigus: if there was an incompatibility in binaries we'd have seen ,... nothing | 14:17 |
krnlyng | rinigus: if there was an incompatibility in java stuff there might be errors later | 14:18 |
krnlyng | rinigus: if there was an incompatibility in the frameworks we'd already have seen different errors | 14:18 |
krnlyng | somewhat like this :) | 14:18 |
rinigus | ok, there are errors on showBoot... and its not optimizing yet. do you want a logcat? | 14:20 |
krnlyng | no it's fine i don't think it will be different | 14:21 |
krnlyng | rinigus: my guess is: surfaceflinger starting too late | 14:22 |
rinigus | krnlyng: maybe. since it seems to be a failure at start | 14:23 |
krnlyng | rinigus: i actually looked in more detail into this error. it is windowmanager service not started up properly thus mWindowManager not set in ActivityServiceManager | 14:23 |
krnlyng | or actually | 14:24 |
krnlyng | window manager service hanging during startup | 14:24 |
rinigus | krnlyng: and my new shiny kernel has some more delays, it seems. | 14:24 |
krnlyng | and that hangs because it cannot connect to surfaceflinger | 14:24 |
krnlyng | hmm! | 14:25 |
krnlyng | and that means surfaceflinger did not register it's service properly? | 14:26 |
rinigus | ps -ef : /system/bin/surfaceflinger | 14:26 |
rinigus | ps -ef : /system/bin/surfaceflinger is running | 14:26 |
krnlyng | try "killall servicemanager zygote main system_server surfaceflinger" | 14:27 |
rinigus | krnlyng: zygote and main were not found. and now? | 14:28 |
krnlyng | rinigus: thats fine | 14:29 |
krnlyng | rinigus: hmm does it behave differntly now? | 14:29 |
rinigus | krnlyng: so far logo only. but maybe I should wait a bit | 14:31 |
krnlyng | hm maybe it's not surfaceflinger it is looking for in this case | 14:35 |
rinigus | krnlyng: it looks similar. logo blinks for some time and then the blinking kind of slows down to still. logcat is full of unhappy messages. for example, "Could not create image space with image file '/system/framework/boot.art'. Attempting to fall back to imageless running. Error was: Only the zygote can create the global boot image" | 14:36 |
krnlyng | maybe Context.DISPLAY_SERVICE is the problem | 14:36 |
rinigus | krnlyng: btw, I think it should not be a problem, but note that I am using multirom over here | 14:36 |
krnlyng | thats from the sfdroid binary | 14:36 |
krnlyng | rinigus: i am using multirom too | 14:37 |
rinigus | ok, good. its a saver when you test things out! | 14:37 |
krnlyng | yes :) | 14:37 |
krnlyng | i wonder if it is related to dumpsys display hanging | 14:38 |
krnlyng | display.qservice* | 14:39 |
rinigus | krnlyng: is it a process? ps? | 14:40 |
krnlyng | it's from qcom/display/msm8960/libqservice/ | 14:41 |
krnlyng | rinigus: can you do /system/bin/dumpsys display | 14:42 |
rinigus | can't find service display | 14:42 |
krnlyng | ah obviously because system_server is crashing ... sorry | 14:43 |
krnlyng | i am going to see if i can reproduce the error | 14:44 |
rinigus | try to use my image :) | 14:45 |
krnlyng | i am :D | 14:45 |
krnlyng | about to* | 14:45 |
rinigus | krnlyng: so, my first contribution to sfdroid is producing a buggy n4 image on which you could test other's problems :) [hopefully, you'll get the same bug] | 14:47 |
krnlyng | haha i don't think it is buggy, i think sfdroid is :D | 14:48 |
rinigus | krnlyng: when you use ballock's image, do you also have to hit enter on sim pin dialog twice? or you don't have sim over there? (unrelated q to sfdroid) | 14:50 |
krnlyng | rinigus: no i don't use this as my main phone, haven't tried a sim with it in 2 months :) | 14:51 |
rinigus | krnlyng: ok, I see. np, I'll try to look into it | 14:51 |
krnlyng | remind me and i will try it after the sfdroid tests | 14:53 |
rinigus | krnlyng: I'll see first if newer ofono could help ... but maybe ballock knows how to fix it. we'll cry for help if its eluding us | 14:54 |
krnlyng | yes i do have the error | 15:07 |
*** rinigus-m <rinigus-m!~sailfish@tabasalu-gw.radionet.ee> has joined #sfdroid | 15:17 | |
rinigus-m | krnlyng: you mean sfdroid error, right? | 15:20 |
krnlyng | yes | 15:21 |
rinigus-m | so, as usual, its reproducible. that's good | 15:23 |
krnlyng | W/dex2oat (13882): Failed to open oat file '/data/dalvik-cache/arm/system@framework@boot.oat' referenced from image /data/dalvik-cache/arm/system@framework@boot.art: Failed to open oat filename for reading: Too many open files | 15:29 |
krnlyng | i think actually this is the problem | 15:29 |
krnlyng | hm | 15:32 |
krnlyng | maybe it is even another error | 15:32 |
krnlyng | ok no | 15:33 |
krnlyng | rinigus: chmod 644 /usr/libexec/sfdroid/system/framework/ext.jar | 15:33 |
krnlyng | rinigus: doesn't solve the problem but shuold probably be done | 15:33 |
rinigus-m | krnlyng: but limits can be increased, if needed | 15:34 |
*** krnlyng <krnlyng!~liar@178.114.22.53.wireless.dyn.drei.com> has quit IRC (Ping timeout: 258 seconds) | 15:38 | |
*** krnlyng <krnlyng!~liar@178.114.51.186.wireless.dyn.drei.com> has joined #sfdroid | 15:51 | |
*** ChanServ sets mode: +o krnlyng | 15:51 | |
krnlyng | rinigus: as far as i can tell it fails to set up /data/data/com.android.providers.settings/databases/settings.db | 16:22 |
krnlyng | rinigus: and thus fails somewhere later when trying to start core services | 16:22 |
rinigus-m | krnlyng: i do have this file, but its 0 bytes | 16:27 |
krnlyng | yes thats the problem i think | 16:28 |
rinigus-m | krnlyng: rm and boot? | 16:31 |
krnlyng | rinigus-m: did not help on my side | 16:32 |
rinigus-m | krnlyng: then i guess its not that. but now when we talk about databases, maybe differences in cm versions are becoming important? | 16:37 |
krnlyng | rinigus-m: yes maybe | 16:38 |
krnlyng | rinigus-m: just for the hell of it, before i dig too deep, can you test a rom with YOG4PA as a base? | 16:40 |
*** ced117 <ced117!~ced117@opensuse/member/ced117> has quit IRC (Ping timeout: 244 seconds) | 16:41 | |
rinigus-m | krnlyng: thats not that easy - i dont have any. ballock is building those and he's on vacation right now | 16:42 |
rinigus-m | krnlyng: what about rebuilding sfdroid for matching base? | 16:43 |
krnlyng | rinigus-m: takes a while | 16:44 |
rinigus-m | :(. maybe i should try to build it during a night, if i get the time. | 16:45 |
rinigus-m | do you guys have build instructions? | 16:46 |
krnlyng | rinigus-m: you don't have the YOG7D sources on your system right now? | 16:46 |
krnlyng | by any chance | 16:47 |
rinigus-m | i do. its just i have to go and would be back only later. i can send messages once in a while, but that would need a bit more ... | 16:48 |
rinigus-m | :) | 16:48 |
krnlyng | rinigus-m: ok it is definately a mismatch | 16:52 |
krnlyng | rinigus-m: i don't even have CMSettings in YOG4PA | 16:52 |
krnlyng | rinigus-m: and that is what causes the break | 16:52 |
krnlyng | rinigus-m: as far as i can tell | 16:52 |
*** ced117 <ced117!~ced117@opensuse/member/ced117> has joined #sfdroid | 16:55 | |
rinigus-m | excellent! then i would just need to rebuild for my base | 16:55 |
krnlyng | rinigus-m: i thought there were only minor differences between those two versions | 17:12 |
krnlyng | apparently i was wrong and wasted a bit of time :( | 17:12 |
rinigus-m | lets see if, after building matching sfdroid, problem will disappear. | 17:16 |
krnlyng | rinigus-m: i am pretty sure it will but to be sure get cyanogenmod sources from build-manifest.xml | 17:17 |
rinigus-m | krnlyng: do you have some readme.txt around with build instructions? | 17:17 |
rinigus-m | yep, i have cm and aosp sources that match sfos image. | 17:17 |
krnlyng | rinigus-m: get /system/etc/build-manifest.xml get cm sources with that cherry pick everything from mako_12 branches, clone dbus, clone sfdroid-inits, run buildscript.sh (https://bpaste.net/show/3bf7cfdcf283, adjust device name!) run pack-source.sh (https://bpaste.net/show/d0182ac14ba3 - but adjust the path to sfdroid-inits), mb2 -t $VENDOR-$DEVICE-armv7hl build with spec: https://bpaste.net/show/07943c62c956 | 17:21 |
ballock | I see you guys have been working hard :) | 17:21 |
krnlyng | rinigus-m: if you have the sources already it should only take ~1 1/2 hours | 17:21 |
krnlyng | ballock: ahoi! | 17:22 |
ballock | rinigus: Gallery reliably crashes with the .secure codec removed on yog4p, so I guess there's more mismatch | 17:22 |
ballock | rinigus: I didn't check the trace yet, nor logcat, though. | 17:23 |
ballock | What do you need to build on yog4p? | 17:23 |
rinigus-m | krnlyng: ok, I'll try that later. since its a build that needs sfos sdk, i would need to organise some hdd space. we'll get back with the feedback on build probably tomorrow evening. | 17:31 |
rinigus-m | ballock: interesting, i do wonder whats else is missing on g4p. | 17:32 |
krnlyng | rinigus-m: you don't have the mer sdk? how did you build sfos then? | 17:33 |
krnlyng | rinigus-m: you can always skip over the rpm and tar the files generated by pack-source and untar on the device | 17:33 |
rinigus-m | ballock: we tried to get sfdroid to work on my hybrid cm image | 17:33 |
ballock | I compiled yog4p on a remote server that I can share if that helps you. | 17:34 |
ballock | You'd need ipv6 access, though, it's on ipv6 only. | 17:35 |
rinigus-m | krnlyng: i have full cm on machine which does not have sfos sdk. sfos was build on the other one (it needs root and there some limitations on hdd mounts). so, I'll try when i get hardware sorted - not a big issue :) | 17:36 |
krnlyng | rinigus-m: ok :) btw if it is too much hassle setting up the mer sdk on the other machine, simply copy the pack-source.sh output to the mer sdk machine :P | 17:37 |
rinigus-m | ballock: but why playback doesn't work on the released cm base? i guess we better move that discussion to other channel. but before that, krnlyng can you try yo record video in sfos with camera and play it on my image? please, just to get n>1 | 17:40 |
rinigus-m | krnlyng: thanks for a tip! | 17:40 |
rinigus-m | ballock: looked again through your messages: i don't think that we need to build anything on yog4p - on that sfdroid works. now i need to understand if i can get it working on my image | 17:50 |
ballock | rinigus: Did I share the yog4p manifest with you? | 17:53 |
ballock | rinigus: otherwise I guess you'd need to do the sfos rebasing yourself | 17:53 |
piggz | krnlyng: pong | 17:57 |
piggz | krnlyng: rinigus-m: my sfdroid is also yog7d | 17:58 |
rinigus-m | ballock: yes, you did share. i think through your android manifest | 18:02 |
ballock | rinigus: ok, good then :) | 18:03 |
krnlyng | piggz: can you send me a logcat of your device? maybe it is the same or a similar issue | 18:03 |
rinigus-m | piggz: my image is based on yog7d as well (to have it close to hybris12). | 18:03 |
piggz | rinigus-m: sure ... | 18:16 |
piggz | krnlyng: will do.... | 18:16 |
piggz | krnlyng: the logcat of auto starting surfaceflinger where it doesnt work, or manually where it does but input is flakey ? | 18:17 |
krnlyng | piggz: where it does not work | 18:19 |
piggz | krnlyng: https://justpaste.it/xbh6 | 18:30 |
krnlyng | hm seems like a different problem than baddea and rinigus | 18:31 |
krnlyng | i will look in more detail later | 18:31 |
piggz | krnlyng: i guess the problem is | 18:32 |
piggz | I/SystemServiceManager( 3786): Starting com.android.server.display.DisplayManagerService | 18:32 |
piggz | W/DisplayEventReceiver( 3786): Failed to initialize display event receiver, status=-19 | 18:32 |
piggz | E/AndroidRuntime( 3786): *** FATAL EXCEPTION IN SYSTEM PROCESS: android.display | 18:32 |
piggz | E/AndroidRuntime( 3786): java.lang.RuntimeException: Failed to initialize display event receiver. status=-19 | 18:32 |
krnlyng | i have seen this before | 18:34 |
piggz | \o/ | 18:36 |
piggz | krnlyng: does kimmoli still have the problem with killing servicemanager? | 18:37 |
krnlyng | but i have no solution as of yet :( | 18:38 |
krnlyng | i think dr_gogeta86 and kimmoli also had this problem | 18:38 |
krnlyng | this is really something with surfaceflinger if i am not mistaken again | 18:38 |
krnlyng | rinigus: sorry for the delay. n=2 | 19:17 |
nh1402[m] | what if n != 2 | 19:40 |
rinigus-m | krnlyng: thank you! | 19:41 |
krnlyng | nh1402[m]: well it's not that bad because x = 5 and pie = 3 | 19:41 |
krnlyng | any ideas why android creates an empty /data/data/com.android.providers.settings/databases/settings.db and then crashes because it is empty? | 19:42 |
*** rinigus-m <rinigus-m!~sailfish@tabasalu-gw.radionet.ee> has quit IRC (Quit: IRC for Sailfish 0.9) | 19:49 | |
nh1402[m] | π = 3.14 | 19:50 |
nh1402[m] | Urm is it trying to populate the database in a different location | 19:51 |
nh1402[m] | It should be that location, but its installed the app in a different location | 19:53 |
nh1402[m] | That's probably why its crashing | 19:53 |
nh1402[m] | Its in droid/exec/data or something but the database is still in /data | 19:53 |
nh1402[m] | Well I think that's the problem | 19:55 |
krnlyng | no i tried putting it into /system and that didnt help | 19:57 |
nh1402[m] | What's "it" | 19:58 |
krnlyng | SettingProviders.apk | 20:01 |
krnlyng | SettingsProvider.apk* | 20:02 |
krnlyng | i guess that is what creates the file, actually i don't know but it seems likely since the folder is com.android.providers.settings/databases | 20:02 |
nh1402[m] | Whose fault is that krnlyng | 20:26 |
krnlyng | not mine | 20:27 |
*** rinigus <rinigus!~rinigus@tabasalu-gw.radionet.ee> has quit IRC (Quit: Leaving) | 21:08 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!