Major improvements to wasm-bindgen-futures

[Pauan]

Sorry I'm late, I struggled with the multi-threading atomic stuff.

This PR contains a few major improvements:

• Code duplication has been removed.

• Everything has been refactored so that the implementation is much easier to understand.

• future_to_promise is now implemented with spawn_local rather than the other way around (this means spawn_local is faster since it doesn't need to create an unneeded Promise).

• Both the single threaded and multi threaded executors have been rewritten from scratch:

  • They only create 1-2 allocations in Rust per Task, and all of the allocations happen when the Task is created.

  • The singlethreaded executor creates 1 Promise per tick, rather than 1 Promise per tick per Task.

  • Both executors do not create Closures during polling, instead all needed Closures are created ahead of time.

  • Both executors now have correct behavior with regard to spurious wakeups and waking up during the call to poll.

  • Both executors cache the Waker so it doesn't need to be recreated all the time.

I believe both executors are now optimal in terms of both Rust and JS performance.

[Pauan]

(The build failures don't seem related to this PR, except for the unix tests, which seem... really weird. It almost seems like a bug with anyref)

[alexcrichton]

Very nice! Does indeed read quite well and seems simple enough!

[Pauan]

(@alexcrichton I fixed all your feedback btw)

[alexcrichton]

I pushed up some stylistic feedback and tweaks that I figured would be easiest to go ahead and do myself. I debugged the issue with anyref and I think we'll need to fix that somehow before this lands. Looks like Promise.resolve wasn't previously used in the wasm-bindgen test suite, and it's now used transitively through wasm-bindgen-futures. The anyref pass means that Promise.resolve is directly imported, as-is, into the wasm module. This is, however, invalid:

> Promise.resolve(1)
Promise { 1 }
> x = Promise.resolve
[Function: resolve]
> x(1)
Thrown:
TypeError: PromiseResolve called on non-object
    at resolve (<anonymous>)

We have to call resolve via dot-notation or somehow set the this parameter, so the wasm-bindgen optimization to directly import Promise.resolve is not actually valid. I'm not sure the best way to fix that at this time.

[Pauan]

I'm not sure the best way to fix that at this time.

Yeah, that's tricky. Can we somehow special-case static_method_of so it doesn't get the optimization? Just as a temporary solution until we get proper WebIDL integration.

[alexcrichton]

Yeah that seems reasonable to me!

[alexcrichton]

@Pauan have you had a chance to take a look at this again? Do you want some help in fixing importing Promise?

[Pauan]

@alexcrichton I did take a look, but the anyref optimizations are over my head, so some help would be appreciated.

[alexcrichton]

Sure thing, I've posted #1795 to fix imports of static_method_of to ensure that this is set correctly.

[alexcrichton]

👍

All green now!