[promise] Spec compliance fixes for async/await

[promise] Async/await edge case spec compliance fix

- Don't read .constructor when returning a Promise from an async function.
  Instead, call out to the internals of Promise.resolve directly.
  This is done by adding back in an "optimization" from an earlier form of
  the async/await code written by Caitlin Potter.
- Async functions always return a new Promise with a distinct identity,
  even if they simply return another Promise.

R=caitp@igalia.com
BUG=v8:4483
Committed: https://crrev.com/7826bfa7893d3890122399088b927b08432e301c
Cr-Commit-Position: refs/heads/master@{#38404}

[Dan Ehrenberg, 2016-08-05 00:10:27]

Description was changed from

==========
[promise] Async/await edge case spec compliance fix

Don't read .constructor when returning a Promise from an async function.
Instead, call out to the internals of Promise.resolve directly.

R=caitp@igalia.com
BUG=v8:4483
==========

to

==========
[promise] Async/await edge case spec compliance fix

Don't read .constructor when returning a Promise from an async function.
Instead, call out to the internals of Promise.resolve directly.
This is done by adding back in an "optimization" from an earlier form of
the async/await code written by Caitlin Potter.

R=caitp@igalia.com
BUG=v8:4483
==========

[Dan Ehrenberg, 2016-08-05 00:10:32]

The CQ bit was checked by littledan@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-05 00:10:34]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-05 00:32:27]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-05 00:32:27]

Dry run: This issue passed the CQ dry run.

[caitp, 2016-08-05 00:43:44]

Test looks good. I do think we should have a similar test for rejections, even
if they pass already

https://codereview.chromium.org/2219623002/diff/1/src/js/promise.js
File src/js/promise.js (right):

https://codereview.chromium.org/2219623002/diff/1/src/js/promise.js#newcode391
src/js/promise.js:391: return value;
looks great. Should we do something similar for PromiseCreateRejected? I haven't
checked

[Dan Ehrenberg, 2016-08-05 01:21:32]

Description was changed from

==========
[promise] Async/await edge case spec compliance fix

Don't read .constructor when returning a Promise from an async function.
Instead, call out to the internals of Promise.resolve directly.
This is done by adding back in an "optimization" from an earlier form of
the async/await code written by Caitlin Potter.

R=caitp@igalia.com
BUG=v8:4483
==========

to

==========
[promise] Async/await edge case spec compliance fix

- Don't read .constructor when returning a Promise from an async function.
  Instead, call out to the internals of Promise.resolve directly.
  This is done by adding back in an "optimization" from an earlier form of
  the async/await code written by Caitlin Potter.
- Async functions always return a new Promise with a distinct identity,
  even if they simply return another Promise.

R=caitp@igalia.com
BUG=v8:4483
==========

[Dan Ehrenberg, 2016-08-05 01:22:36]

Added a similar test for rejections, and fixed a related bug that Sathya and I
found during review.

https://codereview.chromium.org/2219623002/diff/1/src/js/promise.js
File src/js/promise.js (right):

https://codereview.chromium.org/2219623002/diff/1/src/js/promise.js#newcode391
src/js/promise.js:391: return value;
On 2016/08/05 at 00:43:43, caitp wrote:
> looks great. Should we do something similar for PromiseCreateRejected? I
haven't checked

I don't think we need to optimize for the reject case, and it's not needed from
a correctness perspective because rejecting a Promise doesn't collapse/chain in
any particular way.

[Dan Ehrenberg, 2016-08-05 01:23:11]

The CQ bit was checked by littledan@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-05 01:23:21]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-05 01:45:15]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-05 01:45:15]

Dry run: This issue passed the CQ dry run.

[caitp, 2016-08-06 00:55:50]

lgtm

https://codereview.chromium.org/2219623002/diff/40001/src/js/promise.js
File src/js/promise.js (right):

https://codereview.chromium.org/2219623002/diff/40001/src/js/promise.js#newco...
src/js/promise.js:395: function PromiseCastResolved(value) {
nit: an alternative name for this would be very nice, as it is so similar to
PromiseCreateResolved.

Not a blocker

[Dan Ehrenberg, 2016-08-06 00:56:14]

The CQ bit was checked by littledan@chromium.org

[commit-bot: I haz the power, 2016-08-06 00:56:22]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-06 00:58:23]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-08-06 01:00:59]

Description was changed from

==========
[promise] Async/await edge case spec compliance fix

- Don't read .constructor when returning a Promise from an async function.
  Instead, call out to the internals of Promise.resolve directly.
  This is done by adding back in an "optimization" from an earlier form of
  the async/await code written by Caitlin Potter.
- Async functions always return a new Promise with a distinct identity,
  even if they simply return another Promise.

R=caitp@igalia.com
BUG=v8:4483
==========

to

==========
[promise] Async/await edge case spec compliance fix

- Don't read .constructor when returning a Promise from an async function.
  Instead, call out to the internals of Promise.resolve directly.
  This is done by adding back in an "optimization" from an earlier form of
  the async/await code written by Caitlin Potter.
- Async functions always return a new Promise with a distinct identity,
  even if they simply return another Promise.

R=caitp@igalia.com
BUG=v8:4483

Committed: https://crrev.com/7826bfa7893d3890122399088b927b08432e301c
Cr-Commit-Position: refs/heads/master@{#38404}
==========

[commit-bot: I haz the power, 2016-08-06 01:01:00]

Patchset 3 (id:??) landed as
https://crrev.com/7826bfa7893d3890122399088b927b08432e301c
Cr-Commit-Position: refs/heads/master@{#38404}

[gsathya, 2016-11-30 18:02:27]

On 2016/08/06 01:01:00, commit-bot: I haz the power wrote:
> Patchset 3 (id:??) landed as
> https://crrev.com/7826bfa7893d3890122399088b927b08432e301c
> Cr-Commit-Position: refs/heads/master@{#38404}

The test/mjsunit/harmony/async-await-no-constructor.js isn't valid per spec,
because 'await resovled' causes a PromiseResolveThenableJob which calls
Promise.prototype.then which does a constructor lookup.

The test currently works because we skip the PromiseResovleThenableJob in cases
of builtin promises.