Rename *.atomic.wake to *.atomic.notify?

[binji]

JavaScript recently renamed Atomics.wake to Atomics.notify, see tc39/ecma262#1220. Should we do the same for the WebAssembly atomics instructions?

[conrad-watt]

As a counterpoint, OS APIs seem to tend towards "wake", either in name or documentation (wake_on_bit), (interrupting threads) and we've already diverged in convention from JS slightly on wait/wake to be closer to the OS (#36). Caught between two worlds!

[aheejin]

We named those instructions in atomic.wake and atomic.notify in LLVM backend, to be consistent with JS, which of course does not have to be the same with real wasm instruction names though, as long as the binary opcodes are the same. So overall I'm in favor of the change.

[sunfishcode]

(I don't have an opinion on which names, but with @aardappel's work, the LLVM wasm backend has been moving in the direction of harmonizing its text assembly format with the official wat format, so we really should use the same names, whatever they are, since that's the easy part :-)).

[rossberg]

I have a slight preference for keeping wake, since it seems more common, its meaning is a bit more intuitive, and wake/wait makes the aesthetically nicer pair. :)

[aheejin]

I guess we can close it now, after we decided to use *.atomic.notify in the Nov 13 CG meeting.