Avoid using collection as bean type in ManagedBeanTypesTest

[manovotn]

Fixes #476

IIUIC, this should fix the problem while retaining the test goal (which is bean types for nested generics).
Is that right @arjantijms ?

[Ladicek]

I would avoid Iterable too, even if it seems fairly safe.

But yes, this is a proper way to fix this test.

[arjantijms]

Looks good, same remark about Iterable.

[manovotn]

I would avoid Iterable too, even if it seems fairly safe.

Looks good, same remark about Iterable.

Alright, alright, my laziness got the better of me :)
I've updated the PR to only use custom interface structure.

[Ladicek]

Do we treat this (or rather #476) as a challenge?

EDIT: I mean, I don't have a problem with that, and I think we probably should, just asking for opinions. If we don't, this PR should wait for 4.1.

[arjantijms]

It probably can't be a challenge, at least not the way it's worded right now. JDK 21 is obviously not a requirement for the EE 10 TCK.

But just like an implementation detail was updated for the Authentication TCK that didn't effect the test goal itself (which is just testing generics in this case), it could be a general update. Thoughts?

[manovotn]

But just like an implementation detail was updated for the Authentication TCK that didn't effect the test goal itself (which is just testing generics in this case), it could be a general update. Thoughts?

I am not sure we are allowed to put anything but necessary changes and challenge resolutions to 4.0.x releases now.
It's probably safer to leave this for 4.1 - that being said, we should consider just moving main branch to that and creating a separate 4.0 branch for any further maintenance releases. Most work going into CDI now will be related to 4.1 anyway

[Ladicek]

I added the 4.1 milestone.

Good point about branching @manovotn though. I want to submit the PR for TCK tests for InvocationContext.getInterceptorBindings(), and there are some other outstanding PRs for 4.1 already. So I agree with moving master to 4.1.