Daniel Urban

Is it really? All `mapK` does is converting with an user provided `f: F ~> G`, e.g., `take` is `f(self.take)`. If `self.take` is correct (meaning an appropriate `F.uncancelable { poll...

This is a `F ~> G`: ```scala new ~>[F, G] { def apply[A](fa: F[A]) = G.raiseError(new Exception) } ``` It clearly won't work. So the `F ~> G` has to...

> Backward compatible No, it's not binary backward compatible, and mima (rightly) complained.

I think with `generateFunctionSourcePositions(false)` it happens because `hasDebugInfo` is false [here](https://github.com/scala-native/scala-native/blob/bf4736f3cf0e727cd9057e75681bbb84aa741ae1/nativelib/src/main/scala/scala/scalanative/runtime/StackTrace.scala#L201C11-L201C23), so `position` will be invalid. Then, `withNameFromUnwind` uses that position (it doesn't use `ip`) and `cursor`. Previously (before #4512),...

I think limiting the number of fibers to N makes sense. And I think we can preserve (or rather, improve) fairness even with that. I've started working on something like...

@djspiewak Kind of... I've found my old branch: https://github.com/typelevel/cats-effect/compare/series/3.x...durban:cats-effect:parTraverseN_3 (I'm still reading my code, and trying to remember what I was thinking 2 years(!) ago. It's possibly more complicated than...

@djspiewak In my old branch (see above) I've added some tests for parTraverseN; feel free to use them, if they seem useful.

Or possibly just really the windows runners being slow? Because now there is something else: https://github.com/typelevel/cats-effect/actions/runs/16380475513/job/46295378053#step:22:4120

Thanks for the PR. I'm wondering, what is the intention here? To print _which_ thread exactly?

Sure, but which thread, executing what code, when? The thread writing to stdout? The thread executing `this: IO[A]`? The thread executing the closure in the `guaranteeCase`? (Again, I'm asking for...