T42_ | <adampigg> @b100dian any tips for running the coderus container sdk in an action, but using the built-in container support, not running docker manually? can get it to the build step here https://github.com/HelloVolla/sailfish-release-halium-vidofnir/actions/runs/6522834220/job/17712819537 | 08:06 |
---|---|---|
T42_ | <adampigg> i think im mixing up steps to run outside the container, however i dont know how to checkout files on the host, then run steps inside the docker image | 08:29 |
T42_ | <b100dian> I won't be able to look until in the evening (on a computet) but I remember there was a docker parameter mapping a folder to the outer environment, I think I used it to collect the mic output | 08:39 |
T42_ | <b100dian> *computer | 08:40 |
T42_ | <b100dian> Yeah, I think that `container:` runs everything in docker, maybe that's why even checkout doesn't work. Maybe there's a way to use a separate job for the container and call between them? | 08:51 |
T42_ | <b100dian> What was wrong with plain docker command as a step @adampigg :)? | 09:15 |
T42_ | <adampigg> i may go back to that ... i was having disk space issues with my image creation (re @b100dian: What was wrong with ...) | 09:42 |
T42 | <b100dian> So, my understanding is that you need to define separate jobs and make them depend one one another, and only one would be using the container stuff https://docs.github.com/en/actions/using-jobs/using-jobs-in-a-workflow#example-requiring-successful-dependent-jobs | 12:28 |
T42 | <b100dian> mal: trying to build qtmm: "Use of uninitialized value $directory in concatenation (.) or string at /usr/lib64/qt5/bin/syncqt.pl line 685." https://pastebin.ubuntu.com/p/cG6b7bDd4g/ | 18:54 |
mal | ah, that is the issue I have seen before, trying to remember what causes it | 18:56 |
T42 | <b100dian> does it matter if the target is 4.4.0.72 :D? | 19:00 |
T42 | <b100dian> The line is the first one in the while(1) of locateSyncProfile | 19:01 |
simonschmeisser | Hi there | 19:01 |
simonschmeisser | mal: do you have some ideas for how to fix the color mismatch in camera2 branch of droidmedia? | 19:02 |
T42 | <b100dian> Mal this is the line. Do I need a sync.profile ..? https://pastebin.ubuntu.com/p/krKJDnRp5B/#line-number-685 | 19:05 |
T42 | <b100dian> simonschmeisser: I'm looking into this too. My assumption is that this is closer to the presenting code in qtmultimedia. What device are you on? | 19:05 |
mal | simonschmeisser: no ideas yet for that, a strange issue | 19:06 |
mal | simonschmeisser: I was planning on adding some debug printing to qtmultimedia to see what it sees | 19:06 |
mal | @b100dian just to be sure, you have latest platform sdk? | 19:08 |
T42 | <b100dian> Probably not, if I have a 4.4 target | 19:08 |
T42 | <b100dian> Let me look up the mantra to update | 19:08 |
mal | also make sure you have latest tooling | 19:11 |
mal | "sdk-foreach-su -ly ssu re 4.5.0.24" then "sdk-foreach-su -ly zypper ref" and finally "sdk-foreach-su -ly zypper dup" | 19:11 |
mal | that is how I do it | 19:11 |
T42 | <b100dian> Yeah, this is what I looked up, found a previous one. Would take a while with foreach ig | 19:13 |
simonschmeisser | how does qtmm know the format? changing droid_media_camera_get_video_color_format does not seem to have any effect | 19:13 |
mal | it's possible to update each part manually also if needed | 19:13 |
T42 | <b100dian> my understanding: it goes through gst-droid first. qtmm has a gstreamer plugin | 19:18 |
simonschmeisser | would be lovely to have a readme.md | 19:36 |
T42 | <b100dian> This is high level https://docs.sailfishos.org/Reference/Core_Areas_and_APIs/Multimedia/ | 19:43 |
T42 | <b100dian> But all the "bits" are huge frameworks to dive into, Qt, Gst, Android.. :) | 19:44 |
mal | yeah, camera goes through droidmedia->gst-droid->qtmm | 19:53 |
simonschmeisser | is the preview rendered by qt or by droideglsink? | 19:58 |
mal | not quite sure | 19:59 |
simonschmeisser | hehe | 20:01 |
simonschmeisser | gst-inspect-1.0 droidcamsrc gives one assert and a segfault near the end | 20:01 |
T42 | <b100dian> checking logs here, it should work with `gst-inspect-1.0 --gst-disable-segtrap --gst-disable-registry-fork droidcamsrc` | 20:13 |
T42 | <b100dian> ah, it segfaults also here after some page full of output | 20:15 |
T42 | <b100dian> simonschmeisser wanna gdb —args it?:) | 20:19 |
simonschmeisser | no gdb by default on sailfishos ... I thought it's a hackers os :-D | 20:24 |
T42 | <b100dian> Just `zypper in` it along with gstreamer1.0-debug* packages | 20:28 |
T42 | <adampigg> @b100dian So, i went back to the original method .. if you look in the log, there is only 4gb of space .. any idea how to get more? https://github.com/HelloVolla/sailfish-release-halium-vidofnir/actions/runs/6526284623/job/17719867558 | 20:28 |
simonschmeisser | will try tomm | 20:32 |
simonschmeisser | tomorrow | 20:32 |
mal | camera2 code likely has many bugs | 20:34 |
T42 | <b100dian> @adampigg maybe https://docs.docker.com/engine/reference/commandline/run/#storage-opt ? | 20:46 |
T42 | <b100dian> mal: got the same sync.profile error with 4.5.0.24 tooling/target | 20:48 |
T42 | <adampigg> @b100dian not sure ... i print that before the container, so its the host size https://github.com/HelloVolla/sailfish-release-halium-vidofnir/blob/master/scripts/create-image.sh#L7 | 20:48 |
T42 | <b100dian> ah, gotcha | 20:49 |
T42 | <b100dian> maybe add a `df -h` before docker run around build.yml#L27 ? It really sounds as if it is space eaten by the build | 20:55 |
T42 | <b100dian> also try `docker run --rm `... | 21:00 |
T42 | <adampigg> trying an action that removes things from the host.... | 21:03 |
T42 | <adampigg> i think with a mix of this, and the docker option, i can figure it out :) | 21:09 |
T42 | <b100dian> mal: I just commented out the "in-source" part https://github.com/sailfishos/qtbase/blob/mer-5.6/bin/syncqt.pl#L702-L708 | 22:10 |
T42 | <b100dian> $out_basedir was something like /srv/mer/targets/$target/usr/share/qt5/mkspecs/features | 22:11 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!