[Test] Move leakwheel to Pri1

[fanyang-mono]

Fixes #53386

GC test leakwheel contains random number generator
https://github.com/dotnet/runtime/blob/main/src/tests/GC/Scenarios/LeakWheel/leakwheel.cs#L158

Half of the time it ran out of time on Android x64. I am moving this test to Pri1, due to its randomness.

[ghost]

Tagging subscribers to this area: @CoffeeFlux
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Fixes #53386

GC test leakwheel contains random number generator
https://github.com/dotnet/runtime/blob/main/src/tests/GC/Scenarios/LeakWheel/leakwheel.cs#L158

Half of the time it ran out of time on Android x64. I am moving this test to Pri1, due to its randomness.

Author:	fanyang-mono
Assignees:	-
Labels:	area-AssemblyLoader-mono
Milestone:	-

[trylek]

LGTM infra-wise. I defer to Maoni to comment on the test as such and its categorization in greater detail.

[trylek]

In particular, in the Teams channel we were discussing what is the most appropriate tagging of the test, whether it's just Pri1 or something more involved like IsLongRunningGCTest. I don't think my GC expertise is sufficient to make a call on that.

[mangod9]

Hey @trylek is it possible to switch the priority only for Mono? Are the tests compiled only once or separately for coreclr and mono?

[trylek]

As I told @fanyang-mono earlier today, it should be theoretically possible to make long-running or other test properties dependent on the test project properties. Having said that, we're still largely struggling with a random library test failing no-one pays initial attention to as it's a networking test and these are supposed to fail once in a while. Deciding on and coding up the support for this behavior is probably the most challenging part.

[mangod9]

Ok, I think this is ok for now.

[fanyang-mono]

Failures on windows arm64 is irrelevant.