Daniel Urban

This should be fixed now (#4426 and #4388, released in 3.6.2).

I'm probably missing something, but do we _really_ need the `ParkedSignal.Interrupting` state? What does the spinlock _actually_ protects? (Spurious wakeups already must be handled.) (Btw, I've restarted the "windows-latest, 2.13.15,...

Okay, so the way to wake someone up is (without this PR) essentially: ```scala if (signal.getAndSet(false)) system.interrupt(...) ``` With the 2 different parking states it could be: ```scala val old...

@djspiewak Yeah, I haven't looked at `doneSleeping` yet, makes sense. Also, I completely understand being cautious. ---- Something else: https://github.com/typelevel/cats-effect/actions/runs/14459016996/job/40548044386?pr=4377#step:22:4843 This is the second time I've seen this branch hang...

Java `static final`s also seem to be inlined at use sites: ```java public final class Foo { public static final long Foo = 42L; } ``` ```scala final class Bar...

Yeah, something is definitely wrong here; possibly more than one thing. What I'm (almost) sure is wrong is `PureFiber#cancel`: https://github.com/typelevel/cats-effect/blob/series/3.x/kernel-testkit/shared/src/main/scala/cats/effect/kernel/testkit/pure.scala#L391. It just writes `Outcome.Canceled()` into an mvar. So if the...

Yeah, that's definitely a bug; either in the code or in the docs.

Was fixed by #4160.

I think it'd have to be a linked structure, not an array. Which is basically SynchronousQueue in unfair mode, as noted above. (I realize it's an implementation detail, but in...

I've now added a test to #4426 (it's not that great, since it's somewhat timing-dependent, and directly inspects thread states, but I didn't have a better idea). That is PR...