[test] Fix flaky unit tests

[cherniavskii]

Trying to fix https://app.circleci.com/pipelines/github/mui/mui-x/78497/workflows/58a06253-0d60-4507-8e42-a6ea37959cce/jobs/446898?invite=true#step-106-8172_107
I wasn't able to reproduce the failure locally (believe me, I tried), but I'm using the same approach as in #16219 that seemed to work.
Hopefully, this will reduce test fails on master and my inbox won't be flooded with false negatives 😅

[mui-bot]

Deploy preview: https://deploy-preview-16231--material-ui-x.netlify.app/

Generated by 🚫 dangerJS against 855a061

[lauri865]

I also tried to locally reproduce this and also other similar issues without any luck. I guess running it in a docker container with constrained resources could be a way to replicate Circle CI, but I don't have time to play around with it myself.

The downside to using waitFor is that it's more forgiving to any performance regressions / delayed state updates, as well as slows down tests (will wait for up to 1s by default). But until anyone finds a good way to reliably reproduce the errors locally, it's a pain to fix in any other way than to use waitFor, unfortunately.

It seems like theses tests started failing more consistently since recently (e.g. the locale test I encountered before, but usually cleared on the next run, but now seen it happen a couple of time in a row). Not sure why that is, but just an observation.

Probably turning off any animations/transitions in a testing environment could help though. Seems like it might be the root cause for both this and column pinning test, as the test itself doesn't seem to be failing, but other components get delayed state updates as a result of user interaction / re-rendering.