Run CI (tests) on MacOS

[joanlopez]

What?

It enables macos-latest for test-related (test.yml) CI workflows.

Why?

Ha, why not? Jokes aside, imho now that MacOS-related issues (#2578 and #1230) have been closed (fixed), and that it looks like MacOS runners have at the very least a performance equivalent to Windows runners' (in fact, test jobs seems to be taking less time in MacOS runners), I guess it's good idea to make the check of the good functioning of k6 on MacOS part of the CI checks.

Also, in the worst case we can disable it again, but why not giving it a try?

Checklist

• I have performed a self-review of my code.
• I have added tests for my changes.
• I have run linter locally (make lint) and all checks pass.
• I have run tests locally (make tests) and all tests pass.
• I have commented on my code, particularly in hard-to-understand areas.

Related PR(s)/Issue(s)

I haven't found any.

[codecov-commenter]

Codecov Report

Attention: 17 lines in your changes are missing coverage. Please review.

Comparison is base (d814297) 72.54% compared to head (1a6d5cc) 72.46%.
Report is 5 commits behind head on master.

❗ Current head 1a6d5cc differs from pull request most recent head 8047e70. Consider uploading reports for the commit 8047e70 to get more accurate results

Files	Patch %	Lines
execution/scheduler.go	57.57%	7 Missing and 7 partials ⚠️
execution/controller.go	75.00%	1 Missing and 1 partial ⚠️
cmd/inspect.go	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3528      +/-   ##
==========================================
- Coverage   72.54%   72.46%   -0.08%     
==========================================
  Files         278      278              
  Lines       20891    20933      +42     
==========================================
+ Hits        15155    15170      +15     
- Misses       4773     4788      +15     
- Partials      963      975      +12     

Flag	Coverage Δ
ubuntu	72.46% <70.17%> (-0.01%)	⬇️
windows	?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[joanlopez]

For some reason (I didn't fully get), without the extra space it doesn't work on MacOS, but with it, it works correctly in all platforms 🤷🏻

[olegbespalov]

I remember that we had issues with codecov (which later were resolved) and created the task, but one of the highlights was that codecov has the GH action. So maybe it's worth migrating to it

[joanlopez]

Oh, nice! Thanks for letting me know, I could consider working on that right after, but I'd prefer to keep it for a separate PR, if that sounds good to you as well 🙏🏻

[joanlopez]

I tried luck without this, and there has been no way to get all tests passing. I guess next step will be to take a look at that flakiness and try to correct it, but for now I guess it's fine, as it has been for Windows so far.

[olegbespalov]

The issue is that tests are flaky on all platforms, primarily GRPC-TLS related https://github.com/grafana/xk6-grpc/issues/39. Which ones do you constantly see?

[joanlopez]

Yeah, there's nothing new, as I think all (or most) of them are documented at #2144.

It's just that it seems that Windows and MacOS runners are way more likely to fail (it's probably a matter of their performance), so I decided to add MacOS into that exception we already had for Windows, to avoid adding more flakiness.