Properly rest mocks (again)

[geoand]

Follows up on: #47181.

This should fix the flaky tests we've been seeing on PerClassSpyTest

P.S. If #47181 is backported, then this needs to be as well

[gsmet]

And we still avoid the memory leaks we were trying to fix?

[gsmet]

Should we drop the mocks from TEST_TO_USED_MOCKS after the test?

[geoand]

Now we are trying to avoid flaky tests, we don't have any new reports of memory leaks

[gsmet]

Yes but my understanding is that we were calling clearInlineMock to avoid memory leaks?

[gsmet]

And I'm not very familiar with this so it might be invalid but I'm surprised we keep a reference to all mocks in a static field of MockitoMocksTracker. Shouldn't we clean the CHM once the test has been executed and we have cleaned up the mocks for this test?

[geoand]

Yes but my understanding is that we were calling clearInlineMock to avoid memory leaks?

I really don't remember.

And I'm not very familiar with this so it might be invalid but I'm surprised we keep a reference to all mocks in a static field of MockitoMocksTracker. Shouldn't we clean the CHM once the test has been executed and we have cleaned up the mocks for this test?

Maybe, but that would require more time than I have this week to figure out

[gsmet]

Let's not merge this right away, I will do some archeology.

[gsmet]

I pushed an additional commit that seems to work, while still clearing the inline mocks.

Not completely sure that I did nested tests right but I added a test and it seems to work.

[geoand]

Thanks a lot for taking a deep look!

[quarkus-bot]

Status for workflow Quarkus CI

This is the status report for running Quarkus CI on commit 7582717.

✅ The latest workflow run for the pull request has completed successfully.

It should be safe to merge provided you have a look at the other checks in the summary.

You can consult the Develocity build scans.