00:05 <mischief> RP: i didn't really find anything conclusive on solving it..

10:39 <rburton> Is anyone else having problems with kas 3.2.1? 3.2 worked fine but 3.2.1 mangles paths and fails to load the right yml files.

11:33 <rburton> JaMa: i can't build webos-image

11:33 <rburton> ERROR: Nothing PROVIDES 'nativesdk-spice' (but virtual:nativesdk:qemu_6.2.0.bb DEPENDS on or otherwise requires it)

11:33 <rburton> nativesdk-spice was skipped: incompatible with host aarch64-webossdk-linux (not in COMPATIBLE_HOST)

11:43 <rburton> JaMa: and there's a master->main needed in tensorflow-lite on the cpuinfo repos

12:40 <rburton> JaMa: and libpbnjson-native fails to compile. awwww, i was hoping to see if the webos image is a reasonable demo

15:29 <landgraf> rburton: There's #webos-ports channel with few active devs. Not sure if all of them are here

16:53 <moto-timo> rburton kas-container 3.2.1 is also failing my gitlabci https://gitlab.com/moto-timo/yp-dev-day_virtualization/-/pipelines/765195025

18:24 <rburton> moto-timo: yes, that!

18:27 <rburton> moto-timo: works locally in a container using kas-container. it appears that the git rev-parse --show-toplevel fails (because it falls back to hg), the question is why. there was a git upgrade, i bet its a security thing warning about something or other. someone needs to get kas to emit the stdout in that failure case, the logic should probably explicitly check for "git cannot find a repo" vs "git reported an error"

18:27 <moto-timo> rburton: it seems like it added an extra `/kas` to the paths? I don't know

18:27 <rburton> Yeah, it can't find the repo root, so uses the current directory (kas) and then you include kas.foo, so it looks up kas/kas/foo

18:29 <moto-timo> rburton: functional, kas 3.2 https://cdn.artifacts.gitlab-static.net/57/5f/575fb5589fd7b3e0a4cd453a0aa13fca45c0f78a96f1f9eeea36ef627123449e/2023_01_31/3687928135/4040653109/job.log?response-content-type=text%2Fplain%3B%20charset%3Dutf-8&response-content-disposition=inline&Expires=1675535924&KeyName=gprd-artifacts-cdn&Signature=e7STCC6k8jXqigNGy-etEsss-6s=

18:29 <rburton> yes, we downgraded this morning :)

18:30 <moto-timo> rburton: fail with extra `/kas` as far as I can tell https://gitlab.com/moto-timo/yp-dev-day_virtualization/-/jobs/3702489856#L26

18:31 <moto-timo> rburton: yeah... I hate having to pin a version of the container just for this... for my silly heartbeat CI test project

18:31 <rburton> yes!

18:37 <rburton> moto-timo: if you have a minute (i'm about to wash and eat) then i bet if you hack kas's repos.py where it calls rev-parse and catch the status!=0 case and print stdout (redirect stderr to stdout first) you'll get an error message. and i'm thinking i'll put money on it being a new security feature in git which doesn't like gitlab's permissions in the runner.

18:38 <rburton> moto-timo: posted my thoughts to the kas list

18:38 <moto-timo> rburton: thanks for the pointer... and for posting on the list.

19:59 <moto-timo> rburton: git had a security update in 3.2.1 build : https://tracker.debian.org/news/1415016/accepted-git-12302-1deb11u1-source-into-stable-security/

23:59 <rburton> moto-timo: yeah that was my hunch but I was in a rush. Good work for chasing that down!