Fix issue with __proto__ when using ES6 object literals

Fix issue with __proto__ when using ES6 object literals

It should be possible to create a concise method with the name
__proto__ without setting the [[Prototype]]. Similarly, property
name shorthands with the name __proto__ should define an own
property.

BUG=v8:3818
LOG=Y
R=adamk, dslomov@chromium.org

Committed: https://crrev.com/bc3b2960e3b7abc16440717f289a430aef904cb1
Cr-Commit-Position: refs/heads/master@{#26172}

[arv (Not doing code reviews), 2015-01-16 22:40:41]

Remove unused function

[arv (Not doing code reviews), 2015-01-16 22:42:22]

PTAL

[adamk, 2015-01-16 23:12:04]

https://codereview.chromium.org/858673002/diff/20001/src/ast.cc
File src/ast.cc (right):

https://codereview.chromium.org/858673002/diff/20001/src/ast.cc#newcode259
src/ast.cc:259: property->set_emit_store(!seen_prototype);
Probably a dumb question, but does one of your test cases exercise this logic? I
don't see any places where you set the same thing twice (though some cases set
both [[Prototype]] and create an instance property named "__proto__").

[arv (Not doing code reviews), 2015-01-17 16:46:38]

On 2015/01/16 23:12:04, adamk wrote:
> https://codereview.chromium.org/858673002/diff/20001/src/ast.cc
> File src/ast.cc (right):
> 
> https://codereview.chromium.org/858673002/diff/20001/src/ast.cc#newcode259
> src/ast.cc:259: property->set_emit_store(!seen_prototype);
> Probably a dumb question, but does one of your test cases exercise this logic?
I
> don't see any places where you set the same thing twice (though some cases set
> both [[Prototype]] and create an instance property named "__proto__").

I pretty much assumed that this case would be covered already. I'll add a test
unless I can find an existing test for this already.

[Dmitry Lomov (no reviews), 2015-01-19 10:55:27]

lgtm if the test that Adam asks for is added.

[arv (Not doing code reviews), 2015-01-20 15:31:14]

Add test with multiple proto fields

[arv (Not doing code reviews), 2015-01-20 15:32:48]

Turns out that we did not have a test for this case before so I added one.

[arv (Not doing code reviews), 2015-01-20 15:42:44]

git rebase

[Dmitry Lomov (no reviews), 2015-01-20 15:43:14]

still lgtm

[arv (Not doing code reviews), 2015-01-20 15:43:18]

The CQ bit was checked by arv@chromium.org

[commit-bot: I haz the power, 2015-01-20 15:44:23]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/858673002/60001

[commit-bot: I haz the power, 2015-01-20 16:31:34]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-01-20 16:31:55]

Patchset 4 (id:??) landed as
https://crrev.com/bc3b2960e3b7abc16440717f289a430aef904cb1
Cr-Commit-Position: refs/heads/master@{#26172}