Are Promises and Microtasks introduced into ES6?
Is there any consensus about event loop abstraction in ES6?
Only that a suitable hook is needed that embedders can implement. Luckily that's simple WRT existing semantics (we are oblivious to future turns at the language level) and doesn't demand that we import the event loop itself into the language. The only requirement is that promises are async WRT the calling code. Delivery at the end of the current turn is allowed. Object.observe will require that we define an order, but until then, hand waving is sufficient.
Make sense. So microtasks won't be introduced yet, but Promises and asynchronous execution semantics (enough for Promises) are introduced into ES6. Is it correct?
Yusuke SUZUKI wrote:
Make sense. So microtasks won't be introduced yet, but Promises and asynchronous execution semantics (enough for Promises) are introduced into ES6. Is it correct?
And enough for modules.
And really, as dherman points out, the event loop and shared state concurrency have haunted ECMA-262 like Banquo's ghost, forever.
But yes, for ES6 we are going to do "as little as possible". (I hope this doesn't end as it did in the movie.)
Yes. The only immediate correctness constraint needed in ES6 and the immediate DOM timeframe is the "empty stack" requirement
Stated at domenic/promises-unwrapping#coercethenablethenable-then as
- Assert: the execution context stack is empty.
Stated at promises-aplus/promises-spec#the-then-method as
- onFulfilled or onRejected must not be called until the execution context stack contains only platform code. [3.1].
citing promises-aplus/promises-spec#notes stating:
- Here "platform code" means engine, environment, and promise implementation code. In practice, this requirement ensures that onFulfilled and onRejected execute asynchronously, after the event loop turn in which then is called, and with a fresh stack. This can be implemented with either a "macro-task" mechanism such as setTimeout or setImmediate, or with a "micro-task" mechanism such as MutationObserver or process.nextTick. Since the promise implementation is considered platform code, it may itself contain a task-scheduling queue or "trampoline" in which the handlers are called.
Some of the underlying issues this addresses are explained at the promises-aplus/promises-spec#139 thread and the previous issue threads it cites.
On Fri, Oct 4, 2013 at 3:39 AM, Brendan Eich <brendan at mozilla.com> wrote:
But yes, for ES6 we are going to do "as little as possible". (I hope this doesn't end as it did in the movie.)
I've got it. It's very interesting progress :)
On Fri, Oct 4, 2013 at 3:46 AM, Mark S. Miller <erights at google.com> wrote:
Yes. The only immediate correctness constraint needed in ES6 and the immediate DOM timeframe is the "empty stack" requirement
Thanks. I understood the meaning of these statements.
I'm very interested in implementing Promises and integrating it to ECMAScript engine (e.g. V8, SpiderMonkey, JSC)
Last night, I saw the meeting notes of Sep TC39 meetings carefully and I was very surprised that it is said that Promises will be introduced into ES6. It refers the microtasks and the event loop. But I can't find details of the event loop consensus. Is there any consensus about event loop abstraction in ES6?