Add Promise polyfill

[arv]

This adds a port of Andreas Rossbergs Promise implementation which is what V8 is based on too.

It uses Yehuda's async/asap function to do the async call. This is done by adding a new npm dependency. Fortunately they are using ES6 modules so it is a simple import.

Since Promises are always async I had to add support for async tests to the feature tests. These are done by a new comment at the top, just like we have done before with other comments.

This new runtime file now requires compilation

[arv]

@johnjbarton

@slightlyoff Promise added for you.

[stefanpenner]

last i checked the v8 implementation was lagging behind the spec. We should likely wait till it is improved (or improved) or look to another implementation.

A faulty promise implementation will cause nothing but pain and agony, if we can prevent that that would be the best.

[arv]

The main reason for doing this is to get rid of Deferred in favor of Promise. These are already compatible since both are 'thenable'. I have a followup patch that does this.

I could remove everything but the constructor and then if that would make you feel happier?

[stefanpenner]

@arv no don't doit to make me happy, I was merely "bringing a warning", do with it as you please.

My recommendation would be to grab whichever one suites your needs from the list of compliant implementations: http://promisesaplus.com/implementations

[arv]

@stefanpenner It is a good point that V8's Promise is not A+ compatible at this point.

I think this is better than the Deferred we have so I still want to commit this. I'd be happy to work on fixing V8's promise as needed.

[arv]

I updated the Promise implementation. It now passes almost all the Promises/A+ tests. It fails some Promise.all test because of the implementation does not use an iterator at this point.

[stefanpenner]

@arv awesome!

[johnjbarton]

'This' confuses me, because I don't see a string constant as an obvious but inferior alternative to WeakMap.

[arv]

I don't want to depend on weak maps just yet so instead we use an expando.

[stefanpenner]

i think the weak map stuff might get dropped for now.

(cc @domenic)

[domenic]

Yeah, I killed it recently; sorry about the churn :(. domenic/promises-unwrapping@65a5e62 See also domenic/promises-unwrapping#79 where a possible replacement is being discussed.

[arv]

Great. I'll update.

[johnjbarton]

lgtm, two questions.

[domenic]

This is out of context, so I might not be understanding what your comment means, but you can definitely test for callability: typeof then === "function".

[slightlyoff]

Functionness != callability. But it'll have to do.
On 9 Dec 2013 08:10, "Domenic Denicola" notifications@github.com wrote:

In src/runtime/polyfills/Promise.js:

•    chain(y, deferred.resolve, deferred.reject);
  

•  else
  

•    deferred.resolve(y);
  

• } catch (e) {

•  deferred.reject(e);
  

• }
• });
  +}

+// This should really be a WeakMap.
+var thenableSymbol = '@@thenable';
+
+function promiseCoerce(constructor, x) {

• if (isPromise(x)) {
• return x;
• } else if (x && 'then' in Object(x)) { // can't test for callable

This is out of context, so I might not be understanding what your comment
means, but you can definitely test for callability: typeof then ===
"function".

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/520/files#r8201363
.

[domenic]

It actually is, if you look at the definition of typeof and IsCallable in the spec. Both are true if and only if the object in question has a [[Call]] internal method.

[arv]

I based this on https://github.com/rossberg-chromium/js-promise/blob/master/promise.js and fixed some obvious bugs. I guess I missed this one.