Conrad Watt

Oh, wrt the new OP edit > Memory's `buffer` attribute will detach any `ResizableArrayBuffer` buffers vended by the `resizableBuffer` attribute. The inner bullet just after that line says that the...

Makes sense!

We could probably make `2^53` (or lower) the new Web limit for (64-bit) Wasm memories (ref https://github.com/WebAssembly/spec/issues/1292#issuecomment-790180524).

@littledan instead of an accessor property, what about an explicit method like `.getResizableBuffer()` which detaches the buffer(s) linked to the `.buffer` attribute? To preserve @syg's point that an application may...

Just as a drive-by, maybe the change in https://github.com/WebAssembly/binaryen/pull/4567 is related? Presumably before this change the test wasn't actually running. EDIT: although clearly I've fallen into a time-hole because I...

@juj out of curiosity, given the original (very sensible) motivation for the change, is the [5.5M LOC example](https://github.com/WebAssembly/binaryen/pull/4567#issuecomment-1086068895) behaving better now that it's in place? Presumably it involves more than...

I believe, based on the discussion in https://github.com/WebAssembly/threads/issues/7, that it's possible to implement `i64.atomic.wait` as a 64-bit atomic read (to confirm the expected value), conditionally followed by a linux 32-bit...

Ah, true. Sorry for the rookie mistake.

> Of course, this all relies on having the write (sub)typing rules. In particular, if heapref $h

This is an apples-to-oranges comparison. There's a compounding cost to adding another dimension to the type system in that it touches the design and implementation (EDIT: and tooling) of every...