Saturday, 2022-04-09

*** ggabriel is now known as Guest154008:38
T42<kquote03> Hello, is up to date?10:08
T42<kquote03> if not is there a more current place to find other ports?10:10
T42<adampigg> its probably not up to date, but probably the most up to date10:11
T42<kquote03> I see, thanks (re @adampigg: its probably not up ...)10:12
*** ggabriel is now known as Guest154410:51
T42<Inacent> Hi12:10
T42<Inacent> Is that how it should be? :
T42<Mister_Magister> clearly not12:14
voidanix[m]@Spidey24Z about the test suites error: doing rm -rf droid-local-repo repeatedly until it builds works12:23
voidanix[m]absolutely no idea why it does that but i do get the same error12:23
voidanix[m]that's when building droid-hal anyways12:24
T42<kquote03> Hello, I'm thinking of going to hybris-18.1, the pastebin links on don't seem to work, so I'm assuming I no longer need to apply them.12:33
T42<kquote03> Do I need to change anything else like in ?12:33
T42<kquote03> For android_version_major do i need to set that to 11 or should I keep it at 10?12:34
T42<elros34> so you do not debug anything you just switch hybris-16->17->18->? obviously for android 11 based adaptation you cant use android_version_major 1012:40
T42<kquote03> I don't have the skills to debug the display driver, almost all of this process has been beyond my depth if I'm being honest. 9/10 times i just operate on having a hunch12:52
T42<kquote03> I saw that there were substantial changes in LOS 18.1 regarding the display driver, so right now this is the best 'lead' i have (re @elros34: so you do not debug ...)12:52
T42<kquote03> please forgive me for my stupid questions (and any I will ask in the future), cuz even though there's been little progress, I've learnt alot from this12:54
T42<elros34> so similar to others, including me but still you can sometimes have luck and find somebody here who will know how to help you12:55
T42<kquote03> I have actually, @/HengYeDev , he seems to be running into the same issue12:55
T42<kquote03> but on a slightly different SoC12:56
T42<elros34> "could not find or load the Qt platform plugin "xcb0"" from logs @Spidey24Z posted does looks like some silly mistake, we do not use xcb12:56
T42<kquote03> I don't seem to have this issue, I will contact Ayman. Probably a silly mistake like some environment variables.13:14
T42<kquote03> These are the logs from my end13:14
T42<kquote03> dmesg:
T42<kquote03> journalctl:
T42<kquote03> logcat:
T42<kquote03> I tried masking user@100000 and running minimer manually, but I immediately get this (re @kquote03: Hello, https://paste...)13:16
T42<kquote03> we currently think that it's decon causing our issue13:20
T42<kquote03> I tried cat  /sys/fs/pstore/dmesg-ramoops-0 but it prints gibberish (probably encrypted as it had a .enc extension once)13:20
T42<kquote03> I did not have a console-ramoops-0 as @\HengYeDev13:20
T42<kquote03> As for pmsg-ramoops-0 :
T42<elros34> I do not get why do you grab pstore, does your device reboots?13:24
T42<kquote03> nope, i just blindly copied some steps (re @elros34: I do not get why do ...)13:25
T42<elros34> then use dmesg (-w for live log)13:25
T42<kquote03> Alright13:25
T42<elros34> so exynos and decon driver, I guess you tried mce fb control fix?13:26
T42<kquote03> only thing I've tried was masking mce and running it after the system initialized, but I have not tried any fix (re @elros34: so exynos and decon ...)13:26
T42<kquote03> dmesg (right after minimer crash)
T42<elros34> I was talking about this: IIRC clear indication this is needed was working act dead mode but failing normal boot13:32
T42<kquote03> Act dead mode? You mean booting sailfish after the charging indicator13:37
T42<kquote03> Sadly this didn’t change anything, I will give it a try nonetheless (re @elros34: I was talking about ...)13:37
T42<kquote03> Act dead mode? You mean booting sailfish after the charging indicator ?13:37
T42<kquote03> Sadly this didn’t change anything, I will give it a try nonetheless (edited) (re @elros34: I was talking about ...)13:37
T42<elros34> not really, just booting to sfos charging screen should work.13:38
riniguswhat was the point of do we have to go through that release before moving to 4.4.0 or I can directly bump the port to 4.4.0 from
T42<elros34> changelog is short but for me most important is fix for ssu cache bug which I think you even reported once14:00
rinigus@elros34: where did you see the changelog for it?14:03
rinigusfound it!14:03
rinigushmm, looking at it - still not clear whether I can skip it while updating to 4.4.0 ...14:05
T42<elros34> if you skip it then you may have not enabled adaptation-common repo during -> 4.4.0 so some packages will be in incorrect version14:07
Thaodanrinigus: was the polkit hotfix14:08
Thaodanhis supplemental Suomenlinna release contains the fix for the recently found vulnerability in the polkit package (CVE-2021-4034) and it avoids some potential problems when installing the upcoming 4.4.0 release. Improvements for SDK.14:08
rinigus@elros34: did they changed something in adaptation-common that is very much different from what we had earlier. note that tama repos have different url scheme anyway, so I may not see that change in tama14:09
rinigusThaodan: "avoids some potential problems" is rather vague. anything to watch after?14:10
T42<elros34> be aware I am talking about new adaptation-common repo from jolla which contains for example gst-droid not adaptation-community* repos. I am not aware of any other important change, just fix for missing repo14:12
rinigushmm, but as changelog doesn't have any of port packages, it should be trivial then to provide
rinigusjust have to copu all from and it should be fine14:12
T42<elros34> most likely14:13
riniguswith build, I am getting "gnu-which conflicts with busybox-symlinks-which". not sure how to work around it at
T42<elros34> probably in "Project Config"14:42
rinigusdid try to set gnu-which as preferred - didn't help14:42
T42<nephros> I worked around that with #!BuildIgnore in the spec.14:44
rinigusnephros: do you have somewhere an example?14:47
T42<nephros> Correction: tried to but the obs builder then cant install it.14:48
T42<nephros> Maybe a Prefer: helps, checking...14:48
T42<nephros> rinigus:14:49
rinigus@nephros: yes, it builds on 4.4.0, but not on seems the same as you have14:51
T42<elros34> maybe missing similar conflicts with "gnu-which" package: how about adding it to droid-hal spec?14:56
rinigus@elros34: I don't see how it will help as we pull it in build time15:02
rinigusinteresting that busybox which is not installed in 4.4.0 build15:04
T42<elros34> now both packages are installed but there should be only one: busybox or gnu15:04
rinigusexactly. and I have no idea why15:05
T42<elros34> because missing "Conflicts: gnu-which" in busybox-symlinks-which15:07
rinigushow did then others managed to sneak around it? I can see mal  and piggz have their .15 built without any issues. or was something in config changed later?15:09
T42<Spidey24Z> Fixed by adding %define android_config\ (re @SailfishFreenodeIRCBridgeBot: <voidanix[m]>@Spidey...)15:14
T42<Spidey24Z> Have no idea (re @elros34: "could not find or l...)15:15
voidanix[m]hmm, always had that though15:19
malrinigus: what issue are you talking about?16:07
T42<nephros> I'm using sfbootstrap to make packages, and at hte end it sais: ">> Skipping build of droidmedia and supporting packages due to missing HAL bits", which then later leads to: "Warning[04/09 16:10:46] : repo problem: nothing provides 'droidmedia' needed by the to be installed gstreamer1.0-droid-0.20210820.0-1.2.1.jolla.aarch64,".16:11
T42<nephros> So, which are the "missing HAL bits"?16:11
malyou probably didn't run "make droidmedia" in the same environment you run "make hybris-hal"16:13
T42<nephros> mal: did that, and it looks like it completes successfully.16:15
T42<elros34> do you have maybe in system/lib64 directory:
T42<nephros> sfbootstrap build hal output:
T42<nephros> Elros: yes, I have both16:20
T42<nephros> out/target/product/tuba/system/lib/  out/target/product/tuba/system/lib64/libdroidmedia.so16:20
T42<elros34> clearly  [ -f "$ANDROID_PRODUCT_OUT/system/lib/" ] fails so either variable is wrong or file doesn't exist16:24
T42<nephros> makes sense. I'll try hacking the script. Thanks for the hint.16:38
T42<elros34> run script with bash -x <script> or add some "ls -al" before that "if" check16:40
T42<nephros> I shall soon. Right now my buildenv has died - /var/lib/rpm is empty :O16:42
rinigusmal: failing to build
T42<nephros> Elros: found it: $DEVICE != $HABUILD_DEVICE so it looks in the wrong path.16:58
malrinigus: check it now17:06
rinigusmal: thanks! let's see17:08
malseems some issues with the build, trying get a log17:08
rinigusmal: it ends on "file /usr/bin/which conflicts between attempted installs of gnu-which-2.21+git2-1.2.1.jolla.aarch64 and busybox-symlinks-which-1.33.1+git3-1.7.10.jolla.aarch64"17:09
rinigusbut gets into some weird loop17:09
malrinigus: now it finished17:11
rinigusmal: thank you! what did you do?17:12
malrinigus: fyi what I did was changing the order of the repos in meta17:12
malrinigus: so the most generic should be at the bottom of the list and more specific ones at the top17:12
malthat is the order which those are used17:13
malso it somehow got confused because of the wrong order17:13
rinigusmal: interesting!17:13
mal@nephros could you also change the order of the repos in the common repo should be before sailfishos:latest17:16
T42<nephros> mal: sure thing!17:16
mallike here
T42<nephros> done!17:17
malalso check the 4.3.0 common repo17:18
malit's pointing to latest not 4.3.017:18
malah, but you are using devel for that17:18
malwhich is always latest17:18
T42<nephros> :) yea17:20
T42<nephros> Should I prefer testing?17:21
malif you want to build against some specific version then yes17:21
malrinigus: do you need the latest sensorfw on your device, it needs newer gbinder than what is in 4.3.017:22
T42<nephros> ok thx.17:23
rinigusmal: no, I need to specify the tag for sensorfw. wasn't needed when was building, but now it is updated17:23
rinigusmal: should be OK now17:24
malyes, looks fine17:27
Thaodanrinigus, elros34: Busybox is out of the builder because it pulled in selinux which created a circular dependey on selinux, that is why which was restorted.17:29
Thaodancheck whith osc buildconig where there's still an old project config that pulls in busybox17:30
ThaodanThis was part of the work to boostrapping of systemd that allows us to remove selinux and friends from the builder17:31
rinigusThaodan: testing now and nemo-qml-plugin-systemsettings from upstream works as it should! no more need for custom build, thanks!17:32
ThaodanSee here:
Thaodanrinigus: good to hear :)17:33
T42<elros34> I am not sure I follow, still busybox-symlinks-which do not have "Conflicts: gnu-which" so both packages can be installed at the same time17:39
Thaodanelros34: No they can't both provide which.18:30
T42<nephros> OT, but anything which depends on which is broken anyway. command or type should be used.18:39
T42<nephros> Whoo, it finished creating rootfs! Now I only have to create the base, too... :/18:43
voidanix[m]is ofono-binder-plugin supposed to be compilable?
stoefelzhey guys, maybe someone could give me some advice19:13
stoefelzi am trying to build a sailfish img and i stuck at the 5.4 point in the hadk.19:13
stoefelzwhen i start the command 'make -j$(nproc --all) hybris-hal droidmedia' it is working but the process is consuming my whole RAM (32 GB + 8 GB Swap) and the system is then killing the process. I've tried it with -j1 but this doesn't affect the ram usage :(19:13
malvoidanix[m]: if you are trying to build the very latest version then you might need to update some other packages first, maybe just try to build a bit older version19:19
T42<elros34> stoefelz: which hybris version?19:25
voidanix[m]mal: thanks! using tag 1.0.3 works, the dependencies were there and correct which got me confused19:26
T42<nephros> Again being an inquisitive idiot: I can't build a plain lineageos base using the hadk can I.19:44
T42<nephros> ?19:44
voidanix[m]looking at the hybris-18.1 manifest, I think you can but won't have any apps19:55
malthe manifest has been minimized to some extent and it's likely it won't build complete android20:16

Generated by 2.17.1 by Marius Gedminas - find it at!