Sunday, 2019-09-08

T42pVNghiaQT was added by: pVNghiaQT07:03
T42<DylanVanAssche> I built `v4l-utils` which provides `libv4l` for `ffmpeg`. However, when I try to build ffmpeg now, I get `nothing provides pkgconfig(libv4l)`. Did I missed something? Packages are here:
T42<adampigg> @DylanVanAssche tried adding v4l-utils-devel specifically in the .spec?09:18
T42<elros34> As you can see here: your libv4l-devel does not provide pkgconfig(libv4l) but pkgconfig(libv4l1) and pkgconfig(libv4l09:19
T42<adampigg> good job @elros34 :)09:19
T42<DylanVanAssche> Oh my bad :p Will try as soon as I'm back home :)09:41
T42<DylanVanAssche> Thanks!09:41
T42<DylanVanAssche> @elros34 That fixed my issue indeed :D12:26
T42<DylanVanAssche> Thanks!12:26
T42<DylanVanAssche> Are there some issues with OBS?14:44
T42<DylanVanAssche> I want to build gst-plugins-good with v4l2src support but I get the following message: `400 remote error: connect  Connection refused` and the build failed14:44
T42<DylanVanAssche> Hmmm other packages are building fine... So it must be something with the `gst-plugins-good` package: ssh://
T42<DylanVanAssche> EDIT:
vknecht@DylanVanAssche: maybe must use anonymous https URL rather than ssh on 2222 port ?15:30
mal@DylanVanAssche oh, you are adding v4l support for those, I had done some testing earlier15:35
malI had all of those built15:35
T42<DylanVanAssche> mal: Oh :O too late :P anyway, I learned something in the process :P15:44
T42<DylanVanAssche> vknecht: HTTPS is used instead of SSH for the submodule15:44
T42<DylanVanAssche> mal: I didn't know you already did that :D15:45
T42<DylanVanAssche> I copied the spec file for `mer-core` and only changed one line in the config (to enable v4l2), so the repo should be okay anyway15:49
mal@DylanVanAssche I actually was doing some testing already quite while ago, v4l-utils spec could be simplifies, if you %setup -q -n %{name}-%{version}/%{name} in %prep you wouldn't need the "cd v4l-utils" parts15:49
malalso some of the v4l-utils/ things file in lists would be gone15:50
mal@DylanVanAssche I do quite meny things I never talk about anywhere15:50
T42<DylanVanAssche> mal: Yeah I had zero experience with RPM packaging before that :P that line could improve my spec much further :P15:51
T42<DylanVanAssche> mal: Oh secret santa :D15:51
maldid you use some existing spec as base for the v4l-utils? usually fedora or opensuse specs are quite good for start15:52
T42<DylanVanAssche> mal: I started from the fedora spec. I knew that trick from the commits on mer-core :P15:53
mal@DylanVanAssche it's good to learn rpm packaging16:01
T42<adampigg> mal: i posted those logs yesterday16:15
T42<DylanVanAssche> mal: True :) you never know when it comes in handy :)16:16
mal@adampigg one of those links doesn't work16:25
T42<adampigg> mal: yeah, the one after it shoud16:25
T42<adampigg> its finding a paste service that works with >500k !16:26
mal@adampigg what did the first log have? I only got dmesg16:26
malor is that journal16:26
T42<adampigg> mal: im sre first was logcat, second was journal ... lemme check16:28
mal@adampigg looks like some services fail with signal 9 so all of those are probably because of the same reason, whatever that reason is16:29
T42<adampigg> mal: the logcat has gone from the first link :/16:30
T42<adampigg> ill get again16:30
mal@adampigg you can always email the logs to me if that would be easier16:31
mal@adampigg are you sure selinux stuff is ok on that device? so many android side failures16:32
T42<adampigg> mal: video, BT, sensors all work16:32
T42<adampigg> so much of android must be working!16:33
T42<adampigg> mal: maybe some more .rc files need selinux stubs loaded?16:33
mal@adampigg looks like everything (even hwservicemanager and servicemanager) fail once or twice at first but then start ok later16:35
T42<adampigg> mal: yes, thats what i thought16:35
malwhat exactly is not working on that device?16:37
malvendor.dataqti seems bad16:37
malyou can try adding selinux_stubs to that16:37
T42<adampigg> mal: logcat
piggz_mal: the thing not working is just that vendor.dataqti service atm .... i wondered if it was related to functionfs ? is it even needed?17:21
T42<DylanVanAssche> mal: Since you have already the v4l stuff, can you maybe provide a spec for gst-plugins-good?17:27
T42<DylanVanAssche> mal: These are the same changes as mine, but OBS doesn't want to build the package...
mal@DylanVanAssche not the same, I have replaced --disable-gst_v4l2 with --enable-gst_v4l2, you removed it17:35
piggz_mal: interesting....think i know why the services fail17:35
malpiggz_: do tell17:36
piggz_mal: i did a systemd-analyse plot to get an svg .....17:36
piggz_the boot order shows that and droid-hal-init are staretd before the overlay mount that is needed for selinux_stubs ... even though overlay-mount should be required by local-fs.target17:37
piggz_mal: now, why it ignores this?17:37
piggz_sh-3.2# ls /lib/systemd/system/
piggz_bind-vendor.service  dev-mtp.mount  overlay-system.service  systemd-remount-fs.service  tmp.mount  var-lib-machines.mount17:38
piggz_lets go back to sysvinit!17:38
T42<DylanVanAssche> mal: Still the same error, do you have it on OBS somewhere?17:40
piggz_lets see if an explicit Before=droid-hal-init.service helps17:44
piggz_nope, that breaks boot with a cyclic dependancy!17:48
piggz_mal: fixed :)17:56
piggz_becuase its an early, local-fs service ... need to add DefaultDependencies=no, to remove a dep on sysinit.target18:01
T42<adampigg> (Document)
piggz_boot still takes for too long thought!18:05
malpiggz_: ah, good, did that help with other errors18:34
malpiggz_: that delay before systemd starts is quite long18:34
mal@DylanVanAssche it did build on OBS but I didn't try on Mer OBS18:38
malmer obs has some issues anyway18:38
T42<DylanVanAssche> mal: Oh that's too bad... It's handy to have it in the repos to install it on the device...18:43
T42<DylanVanAssche> Manually building is the only option then18:43
mal@DylanVanAssche need to investigate why it fails18:47
piggz_mal: it didnt help with the failing dataqti service, but the early logcat look a lot better18:52
malpiggz_: still odd why it's so slow18:53
piggz_mal: line 1531 here is when the boot hands over from the kernel
piggz_it certainly take 44 seconds in kernel loading18:54
malpiggz_: looks like mounting rootfs is very slow18:55
malpiggz_: check the beginning, there is a 10 second gap very early in the boot18:57
piggz_mal: see line 78 ... is the root device specified ok?18:57
T42<DylanVanAssche> mal: Is there anything that I can do from my side? I think you need admin access to Mer OBS to debug these things?18:58
malpiggz_: looks odd18:58
mal@DylanVanAssche not much you can do, I have a feeling it has to do with project configs (not in your project), i486 builds are broken also19:00
piggz_mal: looks like android 9 dm-verity stuff
T42<DylanVanAssche> mal: Oh okay, if you need anything from my side, let me know :D19:00
malfirst thing that needs to be done is get build logs working, without those it's impossible to know what is wrong19:01
T42<DylanVanAssche> mal: Indeed, now we're operating in the blind :(19:02
piggz_mal: in the boot, init-dont.service waits a logn time, what creates the files it waits for? dhi?19:09
malpiggz_: init-done.service should run after user session is finished19:25
*** rubdos_ is now known as rubdos19:40
UmeaboyHi! I'd like to add the i4113 to the Adaptations libhybris wiki page, but I suck at editing the Sony part. Who can help me? The device install page is already created.21:19
r0kk3rzits pretty essy21:35
UmeaboyI want to know exactly where to start copying to NOT make anything twice.21:46
Umeaboy|- gives me a clue, but the it doesn't exist on both sides.21:47
r0kk3rzUmeaboy: fill this out
UmeaboyI couldn't find all the info about the device, but there are my changes.22:56

Generated by 2.17.1 by Marius Gedminas - find it at!