Shu-yu Guo

Thanks for the list! The validations that were implicit in the lexical grammar that must now be made explicit is a very important point.

Interesting example, Jukka, thanks! As @lars-t-hansen says, yeah, there's not really a way to cancel, and the question has baggage around cancellation of Promises. Before I pull on threads to...

> It might be that coordinating workers wrt "letting the SAB die" (what must die is the backing store shared by all workers) is equally complex as what the cancelAllAsyncWaiters...

> I wonder if it would be feasible&sensible to have a new API Atomics.cancel(sabInt32Array[, index]); (Atomics.reject, Atomics.notifyCancel, Atomics.notifyReject?) symmetric to Atomics.notify(), but instead of notifying the waiters, it would make...

> Sorry for probably missing it but: Can someone explain to me why .cancel(sab, index, N) can't be accomplished by passing a signal parameter to waitAsync? Could you spell out...

Thank you for the illustrative example. The uncomfortable but real political answer here is that TC39, where JS is developed, is a different standards body than WHATWG, where `AbortController` is...

> I got the impression that it's important for the cancellation to come from a different Worker (compared to where waitAsync is called), where there's no particular way to share...

> Why? Can you elaborate on this perhaps with a use case where one thread needs the capability of aborting a waitAsync of another thread. @juj lays out the use...

> nobody seems to know how to properly hook up an axiomatic memory model with an operational semantics @rossberg-chromium Are you contending that the 2-stage "filtering" semantics in ECMAScript doesn't...

Andreas, Luke, and I discussed near(er)-term plans for the wasm memory model, focusing on interop with the JS embedding. Please correct me if I incorrectly sum up anything here. The...