Repository navigation
Bug: tests are run inconsistently in CI pipeline #482
Description
Describe the bug
Maybe not a real bug but more an issue I came across when I looked at how our tests are run in CI pipeline:
- core: here we have a matrix covering Python versions from 3.9 - 3.12
- community modules: we have a matrix without real purpose covering only Python version 3.11
So looking at pyproject.toml I guess we want to support Python versions from 3.9 - 4.0, so our pipeline needs to reflect this.
I also see that we do not pin the versions of our Ubuntu runners. I would say it's a better practice to pin version to get predictable CI results.
Also I am scratching my head what purpose that make file still serves there. I see no need to use it in the pipeline jobs.
Also we use coverage, but are not doing anything with that test coverage data.
I could spend some time fixing this up. But I would be very grateful for some feedback and thoughts. Not so sure if I miss something here.
Activity
totallyzen commentedon Mar 19, 2024
hey @max-pfeiffer thanks for the insights!
Personal note: I monitor the testcontainers slack more often than the issues, don't have the time to monitor both, sorry about that - please do come and discuss on slack, (you can also raise the issue here, but it may take me and others longer to act)
Back to your issue:
python-versionfrom that and hard-code it on 3.11make installcan be incredibly useful to set up everything according to our standards (see docs(contributing): add contribution and new-container guide #460)max-pfeiffer commentedon Mar 24, 2024
@totallyzen Ok, so I will create a PR addressing these inconsistency issues and pin runner versions.
I just went through #460 and left some comments and suggestions.
I haven't used slack in ages. But I will check on that channel now.
fix: inconsistent test runs for community modules (#497)