ES6: Implement Object.setPrototypeOf

ES6: Implement Object.setPrototypeOf

http://people.mozilla.org/~jorendorff/es6-draft.html#sec-object.setprototypeof

This just exposes the internal %SetPrototype and adds all the required
type checks as specified.

BUG=v8:2675
LOG=Y
R=dslomov@chromium.org, rossberg@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=18685

[arv (Not doing code reviews), 2014-01-17 17:43:22]

https://codereview.chromium.org/141913002/diff/1/src/v8natives.js
File src/v8natives.js (right):

https://codereview.chromium.org/141913002/diff/1/src/v8natives.js#newcode1020
src/v8natives.js:1020: if (IS_NULL_OR_UNDEFINED(obj) && !IS_UNDETECTABLE(obj)) {
The ES6 spec uses CheckObjectCoercible. It comes up in a few places so maybe we
should add new macro for this?

CHECK_OBJECT_COERCIBLE(obj, 'Object.setPrototypeOf')

which would throw the required type error.

[andrea.giammarchi, 2014-01-17 18:02:43]

Just as "heads up" in case you missed it, I think in line 406 of current master
src/v8natives.js you left the poison in place.

I believe it should not be there so ...

-           set: desc.getSet() === ObjectSetProto ? ObjectPoisonProto
-                                                 : desc.getSet(),
+           set: desc.getSet(),

Even if exposed, the poison was throwing in my previous and incomplete patch I
had to drop that completely in order to make it work as expected.

In case your changes avoid such problem, then never mind this comment.

[arv (Not doing code reviews), 2014-01-17 18:06:38]

On 2014/01/17 18:02:43, andrea.giammarchi wrote:
> Just as "heads up" in case you missed it, I think in line 406 of current
master
> src/v8natives.js you left the poison in place.
> 
> I believe it should not be there so ...
> 
> -           set: desc.getSet() === ObjectSetProto ? ObjectPoisonProto
> -                                                 : desc.getSet(),
> +           set: desc.getSet(),
> 
> Even if exposed, the poison was throwing in my previous and incomplete patch I
> had to drop that completely in order to make it work as expected.
> 
> In case your changes avoid such problem, then never mind this comment.

The proto setter poison pill is orthogonal. I have another patch for it:

https://codereview.chromium.org/103343005/

[Dmitry Lomov (no reviews), 2014-01-17 20:17:36]

Looking good, just agreed with your new macro idea + asking for a bit more tests

https://codereview.chromium.org/141913002/diff/1/src/v8natives.js
File src/v8natives.js (right):

https://codereview.chromium.org/141913002/diff/1/src/v8natives.js#newcode1020
src/v8natives.js:1020: if (IS_NULL_OR_UNDEFINED(obj) && !IS_UNDETECTABLE(obj)) {
On 2014/01/17 17:43:23, arv wrote:
> The ES6 spec uses CheckObjectCoercible. It comes up in a few places so maybe
we
> should add new macro for this?
> 
> CHECK_OBJECT_COERCIBLE(obj, 'Object.setPrototypeOf')
> 
> which would throw the required type error.
 Agreed, good idea.

https://codereview.chromium.org/141913002/diff/1/test/mjsunit/set-prototype-o...
File test/mjsunit/set-prototype-of.js (right):

https://codereview.chromium.org/141913002/diff/1/test/mjsunit/set-prototype-o...
test/mjsunit/set-prototype-of.js:31: function getObjects() {
Maybe more tests to check that slots from new prototype participate in look-up,
while the slots from the old prototype do not

[arv (Not doing code reviews), 2014-01-17 21:03:14]

I added a macro and expanded the tests.

PTAL

https://codereview.chromium.org/141913002/diff/1/test/mjsunit/set-prototype-o...
File test/mjsunit/set-prototype-of.js (right):

https://codereview.chromium.org/141913002/diff/1/test/mjsunit/set-prototype-o...
test/mjsunit/set-prototype-of.js:31: function getObjects() {
On 2014/01/17 20:17:36, Dmitry Lomov (chromium) wrote:
> Maybe more tests to check that slots from new prototype participate in
look-up,
> while the slots from the old prototype do not

Done.

[Dmitry Lomov (no reviews), 2014-01-17 22:40:45]

lgtm

[rossberg, 2014-01-20 10:27:50]

lgtm

[Dmitry Lomov (no reviews), 2014-01-20 10:28:48]

On 2014/01/20 10:27:50, rossberg wrote:
> lgtm

I'll land

[Dmitry Lomov (no reviews), 2014-01-20 10:38:09]

Committed patchset #4 manually as r18685 (presubmit successful).

[Dmitry Lomov (no reviews), 2014-01-20 11:01:15]

On 2014/01/20 10:38:09, Dmitry Lomov (chromium) wrote:
> Committed patchset #4 manually as r18685 (presubmit successful).

Reverted in r18687 for breaking webkit tests:
http://build.chromium.org/p/client.v8/builders/V8%20Linux64/builds/8452/steps...