Sunday, 2021-09-19

T42_<TechTino> ok i checked out a version of the repo from b4 that commit, bcs there were merge conflicts00:02
T42_<TechTino> so should i re-make the kernel etc?00:02
malprobably make hybris-hal00:02
malit should only build things that have changed00:02
mal@TechTino probably would be safer to manually add the changed in the commit back instead of going back in history that much00:03
T42_<TechTino> yeah true00:03
T42_<b100dian> 'e' is a super zoom camera. Jolla-camera needs scrolling for the cameras:) but then maybe just show the ones which were given labels.00:06
mal@b100dian how many back cameras does it have?00:11
T42_<b100dian> 500:11
T42_<XAP2P> wtf (re @b100dian: 5)00:11
malwhat device is that?00:11
T42_<XAP2P> When 8 😂😂00:11
T42_<b100dian> but I remember back in the day that droid-camres was outputting hell00:11
T42_<b100dian> so I am not surprised00:12
malwe probably need to check some debug output to see what it really reports to gst-droid level00:12
mal108 MP?00:12
T42_<XAP2P> a00:13
T42_<XAP2P> aaaaaaa00:13
T42_<b100dian> overselling.. I got it for the sensor size00:13
T42_<XAP2P> I didn't think about the selfie camera : /00:13
T42_<b100dian> 108 is just a gimmick, I make 12Mp photos like everyone else00:14
T42_<b100dian> the point is, the autodetect is off, not that I have 10+ cameras00:15
T42_<b100dian> And I was happy with the main camera access actually (and am)00:15
T42_<b100dian> minus the inability to take videos which I still havent debugged enough00:16
malif you want to check how many it really reports in higher level add something like console.log("num cameras "+QtMultimedia.availableCameras.length); to CaptureView.qml manually, you can see where it uses the availableCameras thing in there00:17
mal@b100dian what is the problem with videos?00:18
T42_<b100dian> mal: videos need more debugging: after 2s they start stuttering in the viewfinder and allocation problems are logged. It seems to me there is noone consuming the frames. Given the (really huge) size of logs I haven't been able to pinpoint it00:19
T42_<b100dian> then it freezes at the end (even with miniaudiopolicy) and so I don't get videos saved00:20
malcan you show what you added to jolla-camera-hw.txt00:21
T42_<b100dian> $ cat /etc/dconf/db/vendor.d/jolla-camera-hw.txt00:22
T42_<b100dian> [apps/jolla-camera]00:22
T42_<b100dian> backCameraLabels=["a", "b", "c", "d", "e"]00:22
maltry that debug printing to the qml file like I mentioned to see what it reports00:23
T42_<b100dian> tried, but this beat me to it: onCameraStatusChanged:510 - num cameras 1500:24
malwow, so it reports 15 cameras00:24
T42_<b100dian> as I said, I remember there used to be some noise from droid-camres on 4.0.1, never worried about it:)00:25
malI know that android can report so called logical cameras which are combining two or more cameras to form a new camera00:26
malfor use in various effects etc00:26
T42_<b100dian> at least the one next to 'e' is locking jolla-camera, lemme reboot and try the others. That would be interesting00:28
T42_<b100dian> tried another one and no avail, I think I need to reboot again00:30
T42_<b100dian> so unfortunately they look like bogus (or miss firmware?)00:30
malnot sure what those are then, some more debug output would help probably00:34
malbut I think jolla-camera needs some fixing to limit the cameras to only those with config00:34
T42_<b100dian> give me a pointer on why droid-camres would not work (e.g. sensor-direction boolean?) I may try again my fork tomorrow
T42_<b100dian> unless you have other means to debug this00:36
T42_<b100dian> agreed, it should also skip unnamed cameras if one likes ["a", "b", "", "d", "e"] only00:36
malyou could also use dump-camera-parameters tool from gtstreamer1.0-droid-tools package00:38
malthat can be used to get all parameters from any camera00:38
malthere ia also this branch for droid-camres can't remember how well that works00:40
T42_<b100dian> so not only 0 or 100:40
malyes, any value can be used in that dump-camera-parameters00:41
T42_<b100dian> ok, this will be interesting. thanks for the time, for now some > 5 params are locking the tool up (just as jolla-camera does), I need the camres, I'll try mine and yours tomorrow00:43
T42_<b100dian> If 15 cameras is your thing you know where you find me:))
T42_<TechTino> ok after a bunch of playing around, and snooping thru the telegram history i think its finally all going smoothly00:55
T42_<TechTino> Forcing installation of 'droidmedia-devel-0.20210902.0-1.noarch' from repository 'Plain RPM files cache'.01:24
T42_<TechTino> '_tmpRPMcache_:droidmedia=0:0.20210902.0-1' not found in package names. Trying capabilities.01:24
T42_<TechTino> No provider of '_tmpRPMcache_:droidmedia=0:0.20210902.0-1' found. man im so close01:24
malyou have too new droidmedia, use these commits for some packages
T42_<TechTino> also, make droidmedia fails with frameworks/av/services/audiopolicy/service/AudioPolicyService.h:34:10: fatal error: 'mediautils/ServiceUtilities.h' file not found01:30
T42_<TechTino> #include <mediautils/ServiceUtilities.h>01:30
T42_<TechTino> so probs related01:30
malyou can look for that ServiceUtilities.h in frameworks folder and fix the file so it find the header, there are some include directory things in there01:31
piggzrinigus: gm07:50
riniguspiggz: morning!08:03
piggzrinigus: found problem in logs with persist...08:03
piggzSep 19 08:47:57 VollaPhone systemd[1]: Starting Mount /home replacement...08:04
piggzSep 19 08:47:57 VollaPhone systemd[798]: mounttmp-home.service: Failed to execute command: No such file or directory08:04
piggzSep 19 08:47:57 VollaPhone systemd[798]: mounttmp-home.service: Failed at step EXEC spawning /usr/libexec/sailfish-device-encryption-community/mount-tmp: No such file or directory08:04
piggzSep 19 08:47:57 VollaPhone systemd[1]: mounttmp-home.service: Main process exited, code=exited, status=203/EXEC08:04
piggzSep 19 08:47:57 VollaPhone systemd[1]: mounttmp-home.service: Failed with result 'exit-code'.08:04
piggzSep 19 08:47:57 VollaPhone systemd[1]: Failed to start Mount /home replacement.08:04
riniguspiggz: just a sec08:05
rinigusforgot to commit08:06
riniguspiggz: done. update sailfish-device-encryption-community08:07
piggzrinigus: i have another thought ... i say in libsfosdevenc, you write .unit files  ... i wondered if that would be better achieved using a systemd-generator?08:07
rinigusI blame 4.2 release at OBS. as my devel was on 4.1 before, I couldn't check packages via OBS08:07
riniguspiggz: as I don't know what is systemd-generator all I can guess that "maybe" :)08:08
riniguslet me search for it08:08
piggzrinigus: its how you write unit files for "dynamic" things .... i use one on the PP to write the boot.mount unit, becuase i dont know if user flashed to SD or emmc08:09
piggzrinigus: this is my simple example, you probably need to do more complex stuff like reading ini files too
riniguspiggz: yes, use of generators does make sense as we could update configs automatically. right now it would be a pain (rerunning of wizard)08:14
rinigusas I can see it could be executable as well - doesn't have to be a script08:15
piggzrinigus: yes, true08:15
rinigus(even recommended in the man)08:15
piggzill take that as a great suggestion by me then08:15
piggzill add as an issue?08:16
rinigusplease add issue (and great suggestion indeed!)08:16
rinigusI wonder where do those dynamic units are supposed to be written?08:17
piggzrinigus: it doesnt matter, i think the path is passsed as an arguement08:19
rinigusyes, looking at that line :)08:20
rinigusok, found in manual: /path/to/generator normal-dir early-dir late-dir . so, called with 3 args08:21
piggzrinigus: yes, and stored in /run...08:21
riniguspiggz: OK. looks like we will need one more exe. I guess generator will use config.ini and device.ini. iff config is ready, it will generate units. otherwise, nothing will happen on that stage08:24
riniguswhich all means that users already exposed to encryption-community would have to either do something manually or you will need some kind of OTA script to wipe the current config units.08:25
piggzrinigus: i dont think that applies to any on my part yet...08:26
rinigusall in all, great suggestion! should help us to make it more stable.08:26
piggzmaybe 1 or 2, and im happy to make them reflash08:26
piggzrinigus: think you missed a line here :)
riniguspiggz: I did. and even script was pushed which is not the last version... let me recover it from device08:42
riniguspiggz: let's see if it works now :)08:45
piggzrinigus: yep, i have actdead working now ... not i'll hold off any updates until we have new generators :D08:53
riniguspiggz: yes, please do. it should be easy to do and I have an idea on how to implement it08:55
piggzrinigus: i expect an update by this evenin:D08:58
riniguspiggz: let's see. may get tight with the time, then tomorrow.08:59
piggzrinigus: hopefully it will make the whole system simpler?08:59
riniguspiggz: it will make whole system simple to update which is the greatest advantage. as when we discover some bug, will be trivial to push new generator and get updated unit files after reboot09:01
rinigusbut unit files will be as they are, I think. which is relatively simple anyway09:01
T42_<TechTino> hey guys, im still having issues installing droidmedia even tho i downgraded to the commit ppl recommend for 4.1 i also updated my sdk to just incase10:38
T42_<elros34> did you build droidmedia by "make droidmedia" in HABUILD? --gg only package it10:41
T42_<TechTino> yeah i did that, i just did it again to be safe10:41
T42_<TechTino> nah still doesnt work :(10:42
T42_<TechTino> same error10:42
T42_<TechTino> make droidmedia goes fine though10:42
T42_<TechTino> output of make droidmedia10:46
T42_<TechTino> ok there might be some stuff in here, do i have to delete the files that r clashing?10:49
T42_<elros34> which files? You can try to install droidmedia-devel manually in target to see what will happen: sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -m sdk-install -R10:55
T42_<TechTino> so its complaining that Problem: nothing provides droidmedia = 0.20210702.0-1 needed by droidmedia-devel-0.20210702.0-1.noarch10:57
T42_<elros34> zypper se -s droidmedia will tell you what local repo provides then. BTW did you export TEMPORARY_DISABLE_PATH_RESTRICTIONS=true, it affect hybris-boot but I wonder whether it's also needed for other things (in your log: Could not determine android architecture. Please pass it as 2nd argument using gettargetar)10:59
T42_<TechTino> oh right i didnt do my most recent build with that11:00
T42_<TechTino> ill try11:00
T42_<TechTino> its only finding droidmedia-devel, theres so reference to droidmedia anywhere, i have it in my external/droidmedia tho11:02
T42_<TechTino> what i noticed is that it tries to use  but this doesnt exist anymore11:03
T42_<TechTino> (during the
T42_<elros34> zypper search only in droid-local-repo/$DEVICE/11:04
T42_<TechTino> no matching items found11:04
T42_<TechTino> wait hold on i see the files here so command might have been typoed somewhere idk11:05
T42_<TechTino> droidmedia-0.20210702.0-1.aarch64.rpm  wait could this be bc its aarch64 and not arm7vhl11:06
T42_<TechTino> ?11:06
T42_<TechTino> my kernel is aarch64 but i set my install to armv7hl so maybe its clashing bcs of that11:07
T42_<elros34> armv7hl? I see you are using aarch64 target11:07
T42_<elros34> at least you named it samsung-starlte-aarch64.default11:08
T42_<TechTino> yeah11:08
T42_<TechTino> but i mean i thought that sailfish os itself is armv7hl so maybe thats why it rejects the package?11:08
T42_<TechTino> while my device is aarch6411:08
T42_<elros34> sailfish support now both 32 and 64 archs so if you choose 64 then use that11:09
T42_<TechTino> ok then idk why its not working :(11:09
T42_<TechTino> can i manually point to an rpm and install it11:09
T42_<TechTino> root@techtino-pc hadk # rpm -i ./droid-local-repo/starlte/droidmedia-localbuild/droidmedia-0.20210702.0-1.aarch64.rpm11:10
T42_<TechTino>   package droidmedia-0.20210702.0-1.aarch64 is intended for a different architecture11:10
T42_<TechTino> so basically i think: make droidmedia is making it as aarch64 and system is rejecting it11:11
T42_<elros34> are you in target? sb2 command you got previously11:11
T42_<TechTino> yeah11:11
T42_<TechTino> in target11:11
T42_<TechTino> [SB2 sdk-install samsung-starlte-aarch64]11:11
T42_<elros34> so make sure you have correct $PORT_ARCH, you used really aarch64 target url11:13
T42_<TechTino> my port_arch has been set to aarch64 for the whole install.11:19
T42_<TechTino> i may just go back and build everything as armv7hl11:19
T42_<TechTino> wait i think i made a mistake assuming its related to ARCH11:21
T42_<elros34> have you checked "Could not determine android architecture. Please pass it as 2nd argument using gettargetarch" I was talking about? file /usr/bin/bash or path/to/droidmedia*rpm in target11:22
T42_<TechTino> bcs i ran rpm command which would assume x86_64 as its my host so nvm11:22
T42_<TechTino> oh ill check that11:22
T42_<elros34> there is also more verobse zypper log which may tell you something11:22
T42_<TechTino> i really just dont get it. im literally pointing to the rpm yet it still complains11:27
T42_<elros34> what about "file" command out? does "sdk-foreach-su -y ssu re" return 4.2.0?11:31
T42_<TechTino> file droidmedia-0.20210702.0-1.aarch64.rpm11:31
T42_<TechTino> droidmedia-0.20210702.0-1.aarch64.rpm: RPM v3.0 bin ARM11:31
T42_<TechTino> [porter@techtino-pc hadk]$ sdk-foreach-su -y ssu re returns
T42_<TechTino> so do i have to upgrade the sdk?11:32
T42_<elros34> instruction should be in target installation page, next time use sdk-manage command from here to install versioned target and tooling:
T42_<elros34> does rpm -qpi droidmedia*.rpm returns aarch64 for non devel version?11:38
T42_<elros34> and latest "file /usr/bin/bash" which I asked just to be sure you really use 64bit target11:41
T42_<TechTino> yeah it does return aarch64. for the rpm command. for file inside sd2 it says file /usr/bin/bash11:44
T42_<TechTino> /usr/bin/bash: ELF 32-bit LSB executable, ARM, EABI5 version 1 (SYSV), dynamically linked, interpreter /lib/, BuildID[sha1]=f0d966889760c020ab367835eedc17f876e3889a, for GNU/Linux 3.2.0, stripped11:44
T42_<TechTino> so i guess target isnt 64 bit hmm11:44
T42_<TechTino> its possible i installed the armv7hl target and renamed it to aarch64 at some point hmm11:45
T42_<b100dian> mal: this is the output from droid-camres:
T42_<b100dian> this is patch to camera to skip the unlabeled items (condition is the same as for the text Label below)
mal@b100dian was there some camera in the first ones which didn't work?16:50
T42_<b100dian> no, the first five back cameras work16:52
T42_<b100dian> I think they + the front one, account for the 6 cameras that can have resolutions detected16:53
malok, so then the condition could just check if the index if larger than available labels16:54
malI think it's unlikely there are broken cameras in the first ones, the extra logical cameras or something are at the end probably16:55
T42_<b100dian> Indeed, this is works
T42_<b100dian> mal: should I follow up with a forum bug? I can't fix this other than with a patch17:58
riniguspiggz: we have the generator. you would have to add one more project (-generator) and trigger the main package as well as the lib rebuild18:27
T42_<adampigg> rinigus: sure ... i just saw all the tags go in18:29
rinigus@adampigg: you could just remove unit files created by wizard and it should still all work on the next boot18:30
T42_<b100dian> sdk-assistant update says Going to update the tooling [SailfishOS-4.0.1], then the target [xiaomi-tucana-aarch64] - this means that the tooling will have an out of date name, am I right?18:50
T42_<b100dian> There isn't a possibility to have the two targets installed in parallel with the $VENDOR-$DEVICE-$ARCH name so I have to drop the existing one, haven't I?18:51
T42_<FoblaYWelson> Has Sailfish been ported to any xiaomi or Samsung devices?18:52
T42_<b100dian> There's a point to be discussed in the next community meeting, about collecting and consolidating info about the ports. In there you can find several sources to look devices up
mal@b100dian no need for a forum post, I can handle that internally at jolla18:58
T42_<b100dian> mal: thanks! This is different than in other packages' case which can be patched/rebased at each release by the community, that's why I was curious how to handle it19:00
malthat app is closed source19:05
T42_<b100dian> yes, but it also means there's no place to file issues for it specifically19:07
malforum is the normal place but no need for it since I know of the issue already19:09
piggzmal: whats the recommended droidmedia version for 4.2?19:33
malpiggz: 0.20210608.1, just a hint for future: changelog at forum :)19:44
piggzmal: why leave the comfort of my konsole and irc client when I can just ask you/19:52
T42_<b100dian> It just crossed my mind that I need to grep the changelog for all the bits I build on OBS myself before calling it a '4.2'19:52
T42_<b100dian> piggz: make a 'grep the changelogs` service19:53
piggzi can barely keep on top of things is it is!19:57
piggzamazfish is suffering i think19:57
T42_<b100dian> Oh no! Tell me, what 4.2.0 package do you need the version for:)20:12
rinigusI thought going through changelog is a part of the routine :). But agree, porting does take time from apps20:36
kabouikRinigus, slighttly off-topic for this channel: I installed Pure Maps (flatpak) on Droidian and my Gnome settings are set to dark mode, but Pure Maps menu still show up in white. Can this be changed?21:44
kabouik^ rinigus21:44
kabouikSorry i mistyped your nick.21:45
kabouikCurious too if OSM Scout Server can be used with Flatpak, since Flatpaks are isolated22:37
kabouikWell, just got the answer to that second question after searching OSM Scout on Flathub. :>22:40

Generated by 2.17.1 by Marius Gedminas - find it at!