Fix accessors to use holder instead of this

The accessors should get the value from the holder and not from this.

These are all data properties and if they get invoked it means that
they should just return the value of the property from the holder.

BUG=v8:3461
LOG=Y
R=verwaest@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=22575

[Toon Verwaest, 2014-07-23 15:10:37]

lgtm

[arv (Not doing code reviews), 2014-07-23 16:34:34]

Update the rest of the getters

[arv (Not doing code reviews), 2014-07-23 16:40:21]

PTAL

This now passes all the unscopables test I have in 384963002 as well as all the
tests in debug mode (I'll run the webkit tests too now)

Please pay extra attention to StringLengthGetter. Everything else was straight
forward.

[Toon Verwaest, 2014-07-23 16:56:29]

LGTM if you fix the test failure (test/mjsunit/regress/regress-1062422.js)

https://codereview.chromium.org/410923003/diff/20001/src/accessors.cc
File src/accessors.cc (right):

https://codereview.chromium.org/410923003/diff/20001/src/accessors.cc#newcode250
src/accessors.cc:250: ASSERT(value->IsJSObject());
As the failing test indicates, this is wrong since you can put this in the
prototype chain of any other primitive value. Just remove this ASSERT and update
the comment.

https://codereview.chromium.org/410923003/diff/20001/src/accessors.cc#newcode987
src/accessors.cc:987: Handle<Object> result =
Handle<Object>(function->shared()->name(), isolate);
You can just write this as Handle<Object> result(function->shared()->name(),
isolate);

[arv (Not doing code reviews), 2014-07-23 17:01:17]

Remove invalid assert and fix comment

[arv (Not doing code reviews), 2014-07-23 17:03:39]

Updated.

Would you mind landing this for me when you get the chance?

Thanks.

https://codereview.chromium.org/410923003/diff/20001/src/accessors.cc
File src/accessors.cc (right):

https://codereview.chromium.org/410923003/diff/20001/src/accessors.cc#newcode250
src/accessors.cc:250: ASSERT(value->IsJSObject());
On 2014/07/23 16:56:28, Toon Verwaest wrote:
> As the failing test indicates, this is wrong since you can put this in the
> prototype chain of any other primitive value. Just remove this ASSERT and
update
> the comment.

I just saw that too. Removing the assert and updated the comment.

https://codereview.chromium.org/410923003/diff/20001/src/accessors.cc#newcode987
src/accessors.cc:987: Handle<Object> result =
Handle<Object>(function->shared()->name(), isolate);
On 2014/07/23 16:56:28, Toon Verwaest wrote:
> You can just write this as Handle<Object> result(function->shared()->name(),
> isolate);

Done.

[Toon Verwaest, 2014-07-23 20:11:39]

Committed patchset #3 manually as r22575 (presubmit successful).