[primer] Add ansible, try multiprocessing

[jacobtylerwalls]

Home-assistant is h u g e and is the single biggest bottleneck in the primer. We've used it for a few release cycles. I think it's time to rotate something else into the primer. We get bug reports from ansible maintainers, let's try that.

[codecov]

Codecov Report

❌ Patch coverage is 0% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 95.96%. Comparing base (042717c) to head (dcb9472).
⚠️ Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
pylint/testutils/_primer/primer_run_command.py	0.00%	1 Missing ⚠️

❌ Your patch check has failed because the patch coverage (0.00%) is below the target coverage (100.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #10685   +/-   ##
=======================================
  Coverage   95.96%   95.96%           
=======================================
  Files         176      176           
  Lines       19525    19526    +1     
=======================================
+ Hits        18738    18739    +1     
  Misses        787      787           

Files with missing lines	Coverage Δ
pylint/testutils/_primer/primer_run_command.py	30.76% <0.00%> (ø)

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[cdce8p]

Do we use multiple jobs for linting? A normal pylint run on the Home Assistant repo with 2 jobs usually takes around 8min. They do have a lot of checks disabled though but it might still be faster than the current setup. I'd be a bit sad to see it removed from the primer output as it's usually quite good at catching changes.

[Pierre-Sassoulas]

Make sense, and if we break home assistant we'll hear about it 😄 Maybe we can add a flag "lint-in-ci" so we don't lose the configuration we put in place and so we can launch the "full suite" if we want ?

[cdce8p]

Make sense, and if we break home assistant we'll hear about it 😄

Only the really bad issues. They disabled most checkers a while ago in the believe that they are covered by rust and mypy. From my experience not all the disabled ones are but I don't really have the bandwidth to argue for it. Running the primer for it was nevertheless useful as it just overwrites the config and enables all checks.

https://github.com/home-assistant/core/blob/2025.10.3/pyproject.toml#L147-L409

Maybe we can add a flag "lint-in-ci" so we don't lose the configuration we put in place and so we can launch the "full suite" if we want ?

That would be a good option. Maybe trigger on an added label / check if a specific label is preset. That way we could make sure to do a "full" run in case there is a high likelihood of a regression.

[jacobtylerwalls]

Do we use multiple jobs for linting?

I think we avoided this in the past as it was an annoying source of indeterminacy. We might have fixed most of that. We can try it out.

I'll add it back, but I'm still not convinced it's worth the cost. Some times it blows out the output from the other packages by emitting hundreds of similar messages.

[cdce8p]

Some times it blows out the output from the other packages by emitting hundreds of similar messages.

This should have been fixed with #10642 🤞🏻

[Pierre-Sassoulas]

Imo the primer is one of the most useful test we have, maybe it's not such a problem if it's long. Typically on new checks or uncertain fix involving inference we want the maximum information that we can have. We could create a "full primer" github label so we can launch everything but in the general case it's fast ? Or launch the primer only if pylint/checkers or pylint/extensions was modified. Adding ansible is nice in any case.