Wednesday, 2023-07-19

CarbontatedCaffeineHello, I have been getting "!! Error: $DEVICE is undefined. Please run hadk" the past few days and I haven't been able to solve it00:06
CarbontatedCaffeineI have run hadk via habuild but it still doesn't fix the problem00:06
T42<elros34> that is pretty clear: you didn't define  DEVICE env variable. Read hadk pdf instruction how to do this properly, then re-enter platform sdk00:09
CarbontatedCaffeineI did though, it's in there00:10
CarbontatedCaffeineHABUILD_SDK [bullhead] camden@Caffeine:~/hadk$ echo $DEVICE "bullhead"00:11
CarbontatedCaffeineIt seems fine? But not in the platformSDK00:11
T42<elros34> 4.2 Setup the Platform SDK in pdf00:13
CarbontatedCaffeineI did that00:13
maldoes the command hadk work in platform sdk?00:14
T42<elros34> don't tell me you did because clearly you didn't, check files or show them00:14
CarbontatedCaffeineI have never ever made this mistake before, and it's happened twice. I doubt it is my error00:16
T42<elros34> aah this new way, so run 'exec bash' and sfossdk, then echo $DEVICE, do yo use some errors?00:21
T42<elros34> I assume you use latest platform sdk?00:21
CarbontatedCaffeineI don't see any errors, but neither do I see anything coming up when I echo $DEVICE00:22
T42<elros34> what about what mal said: hadk command?00:23
CarbontatedCaffeineSorry mal, I didn't get back to you, no hadk command does not work00:23
CarbontatedCaffeine(in the sdk)00:23
T42<elros34> does not work? no output or error or what00:23
CarbontatedCaffeineI get a command not found00:23
T42<elros34> what about 'ssu re'?00:24
CarbontatedCaffeine"WARNING: ssu.ini does not seem to be writable. Setting values might not work. Device release is currently:"00:24
T42<elros34> what about if you put 'function hadk() { source $HOME/.hadk.env; echo "Env setup for $DEVICE"; }00:26
T42<elros34> hadk' lines in ".mersdk.profile" instead basrc00:26
CarbontatedCaffeineHuh, it's already in there, should I just delete it from bashrc?00:28
T42<elros34> sure, if old file still works00:29
CarbontatedCaffeinestill the same result00:30
T42<elros34> bash -x /srv/sailfishos/sdks/sfossdk/sdk-chroot00:32
T42<elros34> is that fresh platform sdk installation or you were trying to upgrade it?00:34
CarbontatedCaffeine"bash -x /srv/sailfishos/sdks/sfossdk/sdk-chroot" No luck with this00:43
T42<elros34> what do you mean? it should generate hundreds of lines00:45
T42<elros34> did you really added exactly hadk() function I showed or this 'if' too?00:46
T42<elros34> so no hundreds of lines but you have no warning about using .mersdk. show 'cat  ~/.mersdk.profile'00:47
CarbontatedCaffeinefile not found00:48
T42<elros34> you said it there previously so really check00:49
CarbontatedCaffeineI thought you said ".mersdkubu.profile"00:50
CarbontatedCaffeinewhich I do have00:50
T42<elros34> no that is for habuild and you have issue with platform sdk00:51
CarbontatedCaffeineI don't have a .mersdk.profile, sorry for that00:51
T42<elros34> ubu part is for ubuntu chroot/habuild00:51
CarbontatedCaffeineoh okay00:51
CarbontatedCaffeineSo what happened to the mersdk.profile file00:52
CarbontatedCaffeinecause it wasn't created00:52
T42<elros34> it's deprecated but lets se if it works00:53
CarbontatedCaffeinenow it works00:54
CarbontatedCaffeineit wants me to use the "$SAILFISH_SDK" variable now00:55
CarbontatedCaffeinewhich doesn't exist00:56
CarbontatedCaffeineOh well, now it works!00:56
T42<elros34> so nothing when you echo $SAILFISH_SDK in sdk?00:57
CarbontatedCaffeineoh wait00:58
CarbontatedCaffeineyes, sorry00:58
CarbontatedCaffeinebut it doesn't seem to work? I wonder why that is00:58
T42<elros34> you didn't override your 'cat  ~/.bashrc' accidentally?01:04
CarbontatedCaffeineI don't think so. Since I deleted it from it01:06
T42<elros34> I thinks you could add '-x' here to see what is going on or ignore it01:09
T42<elros34> of course after adding correct content to bashrc and removing .mersdk.profile01:10
T42<smatkovi> applying these ~100 patches takes a very long time in the virtualbox emulator. i started yesterday afternoon and today it is at patch 6406:59
T42<smatkovi> maybe it's ebcause i set it to kvm, i thought that's the moste performant one07:00
*** phlixi is now known as Guest640808:01
*** phlixi_ is now known as phlixi08:01
T42<elros34> why would you use virtualbox if you have platform sdk? Even with virtualbox this sounds very strange, after all they are just simple patch files.08:37
T42<smatkovi> you're right i didn't think this through08:48
T42<smatkovi> that sounds like a much better option08:49
T42<smatkovi> i thought for building apps the sailfish sdk is needed, but it makes sense actually to use platform sd08:51
T42<smatkovi> k08:51
T42<smatkovi> yes here it's much faster08:55
T42<elros34> with new target I do not think this will be any different than app sdk08:56
T42<elros34> and certainly without default  cpu/ram limitation08:57
T42<smatkovi> Yes09:24
CarbontatedCaffeineHi, I'm having trouble with a couple of identifiers called "PROP_MSG_GETPROP" and PROP_MSG_LISTPROP11:26
CarbontatedCaffeinebut they are defined in libhybris11:26
CarbontatedCaffeineI will give you some info in a second11:26
maldid you apply patches correctly?11:27
CarbontatedCaffeineCan I ask how to apply the patches again if they haven't, just in case?11:28
CarbontatedCaffeineLast time I tried I got hit with git errors11:28
malfaq says how to apply patches, if you want to unapply those before that run "repo sync -l"11:30
CarbontatedCaffeinewhat does that do?11:30
CarbontatedCaffeineI'm getting "fatal: previous rebase directory .git/rebase-apply still exists but mbox given."11:33
CarbontatedCaffeinewhen I try to re-apply patches11:34
malyou can't apply patches twice11:49
malyou need unapply patches before applying them again11:49
malor you need to check if those have been applied11:49
malyou can for example go to folder $ANDROID_ROOT/bionic and check git log shows hybris patches11:50
CarbontatedCaffeineI'm in a bit of a pickle really, there are some repos that didn't revert11:52
CarbontatedCaffeineSo I applied them incorrectly, what can I do now11:53
malif you get that rebase message you need to go to the folder it happened in and run there git rebase --abort11:57
malor was it "git am --abort"11:57
CarbontatedCaffeineI have done this a couple of times11:57
CarbontatedCaffeinebut I'm still stuck11:58
malyou need to make sure all of patched repos are clean11:58
malyou can see in the hybris-patches/ the subfolders (and their subfolders) where it applies patches, you can manually check those folders11:58
CarbontatedCaffeinethat's going to take a while11:59
CarbontatedCaffeineI could just restart and it'd be quicker11:59
CarbontatedCaffeineat this point11:59
CarbontatedCaffeineOkay, I figured that out12:09
CarbontatedCaffeinerepo sync -l succeeded12:09
CarbontatedCaffeineand I stashed the things that had been patched12:09
CarbontatedCaffeineayy it worked12:11
CarbontatedCaffeineI'm very happy with that, I should change my attitude12:13
CarbontatedCaffeineFor context, I'm currently getting that old source code back up12:21
CarbontatedCaffeinemaking sure it actually still compiles12:21
malthere are only couple of repos it patches12:24
CarbontatedCaffeineI realized that after I actually looked12:25
CarbontatedCaffeineNow all I have to do is remember where I left off, another reason to check up on logs12:38
CarbontatedCaffeineI think I was stuck on needing to mount something12:38
CarbontatedCaffeineit was crucial to get working12:38
CarbontatedCaffeineOh, that's right12:40
CarbontatedCaffeinethe systemd service "droid-hal-init" wasn't working12:40
malwere all partitions mounted correctly12:44
CarbontatedCaffeineThat.. I don't remember12:56
CarbontatedCaffeineI will have to see the situation when I get everything working again12:56
CarbontatedCaffeineor not working cause the first thing it tends to do is bootloop12:57
CarbontatedCaffeinewhich hasn't been solved yet12:57
CarbontatedCaffeinebut can be mitigated12:57
CarbontatedCaffeinecool, it's boot(loop)ing13:23
malimportant things are how soon does it reboot, any telnet available etc13:23
CarbontatedCaffeinePreviously I was able to use a command to stop it from booting further13:24
CarbontatedCaffeineand eventually turn off the service to make it further still13:24
CarbontatedCaffeineafter this, I got stuck and took a break13:25
CarbontatedCaffeineI will assign the device an IP13:25
CarbontatedCaffeineactually, I'm forgetting what I did13:31
T42<elros34> lucky you, you can read channel logs13:32
CarbontatedCaffeinethat's what I'm doing lol13:33
CarbontatedCaffeineI don't understand, it used to work13:46
CarbontatedCaffeineit only comes up as usb013:46
CarbontatedCaffeinenot as enp13:46
CarbontatedCaffeineand it's only there for a split second13:47
CarbontatedCaffeineI swear there was a file I created to stop it from going further, but I can't find it anywhere in the logs13:48
CarbontatedCaffeineI'm too tired to figure this out13:50
CarbontatedCaffeineill do it tomorrow13:50
mal@b100dian have you ever tried setting GECKO_CAMERA_DROID_NO_MEDIA_BUFFER env var to true for testing?15:24
T42<smatkovi> this installing java 7 thing gives me headaches. nothing seems to work properly15:47
malwhy are you installing that?15:48
T42<smatkovi> it is needed for the onyx port15:49
T42<smatkovi> but i'm trying with a dockerfile now15:49
T42<smatkovi> the 3rd one, maybe that works15:49
T42<smatkovi> i finally managed to do it, but now i get one of these again: build/core/ *** device/oppo/common/keyhandler: already defined by device/oneplus/onyx/keyhandler.  Stop.16:16
T42<TheVancedGamer> ah, you can safely nuke java stuff16:17
T42<smatkovi> how can i do that? i deleted device/oppo/common/keyhandler and it gets regenerated16:32
T42<b100dian> mal: with ...NO_MEDIA_BUFFERS=1 I get a crash while reading data for the QOMX_COLOR_FormatYUV420PackedSemiPlanar32m format, I have a question 3 days ago17:09
T42<b100dian> The current pull 111 has problems when changing  resoltion while playing, I will ping when I update it17:10
T42<smatkovi> okay i was stupid. i should set up the ubu-chroot from within sfossk not the host shell17:18
T42<elros34> you changed device?17:25
T42<smatkovi> yes i changed to onyx (oneplus x)17:44
T42<smatkovi> i got a kernel now, but building packages gives me this error
T42<smatkovi> i decided to upgrade the sailfishos version from 3.4 to 4.517:45
T42<elros34> that is major change. I suggest to generate new droid config with up-to-date submodule and than copy/adapt changes18:57
T42<smatkovi> thanks. i think i put it in the wrong directory first, but i regenerated it anyway, the generated droid-config looks different from the one i found on github19:31
T42<smatkovi> now i get this error: Repository 'dir:/home/sebastian/hadk/droid-local-repo/onyx' is invalid.19:35
T42<smatkovi> [~plus-repo-1|dir:/home/sebastian/hadk/droid-local-repo/onyx] Valid metadata not found at specified URL19:35
T42<smatkovi>  - Signature verification failed for repomd.xml19:35
T42<b100dian> Is available on obs? I never remember how to tell..20:34
mal@b100dian if you mean testing projects there then just use the .18 target but name the target as .2120:36
T42<b100dian> mal: I mean develop, but thanks I will try testing on tucana later too20:38
malfor devel you can just use latest which is .18 but from adaptation build point of view the difference is not relevant20:39
mal@b100dian for example I use this for .21
T42<b100dian> Ok, but where will e.g droid-camera 0.1.7 come from?20:57
T42<elros34> @b100dian you simple search for 'sailfishos:<version>' project in obs21:07
mal@b100dian that gecko-camera and most of the other packages come from jolla repos when you build image or update your device, you set the version on device with ssu re21:10
T42<b100dian> Thanks both!21:11
CarbontatedCaffeineI was just looking through the FAQ and I realized I hadn't been building the android part with "TEMPORARY_DISABLE_PATH_RESTRICTIONS=true"23:05
CarbontatedCaffeineI personally have some feedback about information finding or going through the HADK, It's annoying having to check a bunch of official/semi-official sources just to make sure what I'm doing isn't wrong23:07
malyeah, it's not ideal23:24
malbut about that path restrictions, it's possible fix the repos so that the env var is not needed23:25
malI really should do it, was it just cpio causing issues?23:25
CarbontatedCaffeineI had cpio already installed23:30
malyes, but in certain android version it prevents build from using it23:30
CarbontatedCaffeineLineage 18.1/Android 11 I'm building23:31
malthat is what that "TEMPORARY_DISABLE_PATH_RESTRICTIONS=true" fixes23:31
malit makes the build allow using cpio23:31
CarbontatedCaffeineNah I didn't do that23:31
CarbontatedCaffeineI'm guessing 17.1 goes for 18.1 as well?23:32
malyes, if you have problems with cpio then yes23:32
malbuild output should mention it23:32
CarbontatedCaffeineI did not have any problems23:33
CarbontatedCaffeineYesterday I did not set "TEMPORARY_DISABLE_PATH_RESTRICTIONS"23:33
CarbontatedCaffeinebut it was still fine23:33
CarbontatedCaffeinethe reason why I am concerned is "before building android part otherwise hybris-boot.img will not include hybris initramfs"23:33
CarbontatedCaffeineAnd I was wondering if not setting this would stop the phone from booting correctly23:34
malah, 18.1 has the patch already for cpio23:34
CarbontatedCaffeineand it would stop the bootlooping problem23:34
malhow soon does it boot loop23:34
CarbontatedCaffeine> "ah, 18.1 has the patch already for cpio" So the flag doesn't need to be set?23:35
malyeah, no need to set that23:35
maland did you get any telnet up23:35
CarbontatedCaffeine> "how soon does it boot loop" As soon as the notification led turns on (I think it has something to do with the networking stuff).23:36
CarbontatedCaffeineit's very quick23:36
CarbontatedCaffeineIt'll be on for like 1-2 seconds and it'll reboot again23:36
CarbontatedCaffeineI tried to route an IP to USB0 but I was not successful, it'd always keep rebooting23:37
malso you see some interface?23:37
CarbontatedCaffeineVery briefly23:37
malI'm trying to understand at which point it reboots23:37
malwhat does that 1-2 seconds mean? the whole time the device is on or what?23:38
CarbontatedCaffeineI will lay out what happens23:38
CarbontatedCaffeine1. The device will boot to the black screen with the google logo23:39
CarbontatedCaffeine2. It'll stay on for maybe a second or two23:39
CarbontatedCaffeine3. The notification LED will turn on and the laptop will pick up a network for a brief moment23:40
CarbontatedCaffeine4. reboot23:40
malso how long is that in total23:40
CarbontatedCaffeineThe phone is on for around 1.5 seconds I'd say23:40
CarbontatedCaffeinenot long23:40
malwhich telnet port is showing?23:40
malmaybe host dmesg could tell that23:41
CarbontatedCaffeine(sorry, it's around 15 seconds)23:41
CarbontatedCaffeineThe network prompt on my laptop only appears for less than a second23:41
CarbontatedCaffeineI just counted23:41
CarbontatedCaffeine> "which telnet port is showing?" None23:42
CarbontatedCaffeinewell not any that I see23:42
malwhat does host dmesg say, it might show something about Mer Boat loader or something23:42
CarbontatedCaffeineright, give me a sec23:42
malok, so 232323:43
CarbontatedCaffeineThe problem is, is that it's such a small window of time :(23:44
malif there is some way you can access the sfos rootfs via twrp for example, then you should make it stop before starting systemd23:44
CarbontatedCaffeineI can't remember what the file to create was23:45
CarbontatedCaffeineit's stuck in the back of my mind23:45
maluse the source :)
malso you can add such a file via twrp or some other way, the rootfs is in /data/.stowaways/sailfishos/23:47
CarbontatedCaffeineThat part I remember! :)23:48
malthat will for sure stop the reboot, but to debug what is happen after that you probably should mask droid-hal-init23:48
CarbontatedCaffeineIt's currently holding, awesome, I'll just give it an ip23:48
CarbontatedCaffeineI keep getting connection refused23:55
CarbontatedCaffeineputting 2323 on the end got me in23:56
malnow that you are in, mask droid-hal-init with "ln -s /dev/null /etc/systemd/system/droid-hal-init.service" and then use this to continue the boot: echo "continue" >/init-ctl/stdin23:58
maland see if it reboots or not23:59
CarbontatedCaffeine"ln -s /dev/null /etc/systemd/system/droid-hal-init.service"23:59
CarbontatedCaffeineI did this23:59
CarbontatedCaffeineand it worked23:59
CarbontatedCaffeinewe had the same idea23:59
malnow you continued the boot and it's still on?23:59

Generated by 2.17.1 by Marius Gedminas - find it at!