search

OpenAMP / libmetal

An abstraction layer across RTOS, baremetal, and user-space Linux environments
https://www.openampproject.org/
Other
259 stars 172 forks source link

CI: fix timezone and pip3 issues #294

Closed arnopo closed 4 months ago

arnopo commented 4 months ago

The build check fails with following errors:

1) ln: failed to create symbolic link '/etc/localtime': File exists Fix it by removing the link creation

2) error: externally-managed-environment Fix it by adding "--break-system-packages" option to pip3

wmamills commented 4 months ago

Please wait until we agree on the way forward for open-amp and then we/I can update this one also.

wmamills commented 4 months ago

@arnopo are you going to do this one or do you want me to? I had it started but I don't have the change to use venv

arnopo commented 4 months ago

@arnopo are you going to do this one or do you want me to? I had it started but I don't have the change to use venv

Both are ok for me. I'm waiting that we fix open-amp. I did not start libmetal update yet.

arnopo commented 4 months ago

I updated to use ubuntu 24.04 environement I also fix dead link for the freeRTOS lib the checkpatch error is due to the update of the FreeRTOS URL

wmamills commented 4 months ago

I pulled this down and pushed to my main branch to test on github. I get a failure on zephyr. I also see that Linux gets the same failures I was seeing on the desktop testing but github does not count them as an error. My job log is here: https://github.com/wmamills/libmetal/actions/runs/9004301644/job/24736958627

wmamills commented 4 months ago

The zephyr failure appears to be a temp failure of the github web server. I ran the same thing again and it worked OK for zephyr. The linux test failures are still there. I switched to 22.04 for the docker image (on desktop) and they are still there.

wmamills commented 4 months ago

The Linux failures are hard to figure out. Initially they failed on the desktop, now they pass without me changing anything. They still fail on github but this does not cause a problem right now as github workflow ignores the error.

I can't spend any more time on this right now so I suggest accepting this PR as is and we can fixup the Linux failures later (or delete them as they are redundant with the other test action).

FYI: I tested this commit to fix github ignoring the test results. It does allow the workflow to fail but I don't have a fix or explanation of the failure right now. https://github.com/wmamills/libmetal/actions/runs/9020751918

arnopo commented 4 months ago

The Linux failures are hard to figure out. Initially they failed on the desktop, now they pass without me changing anything. They still fail on github but this does not cause a problem right now as github workflow ignores the error.

Yes I suspect that the issue is there since a while. As the workflow don't test the output it ignore the issue.

I can't spend any more time on this right now so I suggest accepting this PR as is and we can fixup the Linux failures later (or delete them as they are redundant with the other test action).

i will try to have a look this week. Ok to treat this in a separate PR/issue

FYI: I tested this commit to fix github ignoring the test results. It does allow the workflow to fail but I don't have a fix or explanation of the failure right now. https://github.com/wmamills/libmetal/actions/runs/9020751918