Fix racy doctests

[K900]

On very fast machines (in my case, a 7950X3D with gen4 storage) it's possible that the script will start running the tests before importlib figures out the files are there, and get a spurious ModuleNotFoundError instead. Explicitly flushing the cache before importing things seems to help it work consistently, at the cost of an extra 0.1s or so of runtime.

Change Summary

Related issue number

Checklist

• Unit tests for the changes exist
• Tests pass on CI and coverage remains at 100%
• Documentation reflects the changes where applicable
• changes/<pull request or issue id>-<github username>.md file added describing change
  (see changes/README.md for details)
• My PR is ready to review, please add a comment including the phrase "please review" to assign reviewers

Selected Reviewer: @PrettyWood

skip change file check

[K900]

Not sure if this is changelog worthy, being just a developer (and maybe distribution) facing change, so please review?

[K900]

Also, I'm not sure if something like this is needed on 2.x, because I couldn't find where the doctest code moved to (if it wasn't removed entirely).

[K900]

Also, do I need to add a changelog entry for this or is there some way to skip the check?

[adriangb]

I don't think this needs a changelog entry. You are editing the V2 files already :)

[K900]

I am not, this is for the 1.10.x-fixes branch.

[K900]

Also I think I figured out how to appease the changelog bot.

[K900]

Can you restart CI? I think github is having a moment...

[samuelcolvin]

Simplest to just push an empty commit.

V2 docs code is tested with pytest-examples in the standard pytest run.

[K900]

It's failing with something that looks unrelated now :(

[samuelcolvin]

Thanks for this, force merged. Well look at fixing the fastapi issue soon.

[K900]

Thanks, appreciate the speedy response!