Yasuhiro Inami

Ah, that's sad 😢

Hi @stephencelis ! Yes, I also agree that overloading another `func ..` will become more trickier. IMO, the existing overload in L211: https://github.com/pointfreeco/swift-case-paths/blob/2a12a39dc20334bbf75ef39142c15a40f521b703/Sources/CasePaths/Operators.swift#L211-L216 is also another form of Optional Kleisli...

Or maybe we can say that this PR's impl is more general that should actually sit in `Prelude`-level than in `CasePath`. Then it will make sense that L211 uses `CasePath`...

P.S. For reference, Rust's `tokio` also has `select!` macro that has similar behavior as `Task.select`, which will cancel other works when first one completes. - https://rust-lang.github.io/async-book/06_multiple_futures/03_select.html Also, here's Swift `async`-version...

@phausler > The behavior you are describing is more like TaskGroup. `TaskGroup` only cancels other tasks when error is received, not when successful completion. (You might be mentioning about the...

> The name select was chosen because it kinda has the same behavior as the C api `select`. Looser sockets in the `FD_SET` don't get closed; they stay open for...

@AndrewSB Thanks for your suggestions along with #1! 😄 For #2, I was formerly implementing exact the same code in inamiy/ReactiveAutomaton@561b137 with the new type `NextProducerGenerator`, but reverted afterward to...

Sorry for late reply (I was on the trip, and still busy for another couple of weeks). Thanks for your idea, I'm still glancing your code though, it looks amazing!...

> I also have concern that it'll be ambiguous to constrain State as a reference-type. Oh, I wasn't aware of `State` only allowing reference type 👀 But we can easily...

2eaf9e7 Changed `Dispatcher`-extracting-key from `State.Type` (implicit) to `Dispatcher.self` (more explicit)