Handlify Runtime_GetPrototype and Runtime_GetOwnProperty.

Handlify Runtime_GetPrototype and Runtime_GetOwnProperty.


R=mstarzinger@chromium.org
BUG=

Committed: https://code.google.com/p/v8/source/detail?r=17062

[Yang, 2013-10-01 13:41:49]

On 2013/10/01 13:41:14, Yang wrote:

First baby steps at handlifying stuff. More to follow.

[Michael Starzinger, 2013-10-01 14:46:18]

LGTM with nits.

https://codereview.chromium.org/25478003/diff/1/src/isolate.h
File src/isolate.h (right):

https://codereview.chromium.org/25478003/diff/1/src/isolate.h#newcode757
src/isolate.h:757: bool MayNamedAccessWrapper(Handle<JSObject> receiver,
At some point we want these functions to be the default implementations, right?
So I thing we should overload the name, otherwise we need to rename all
call-sites to drop the "wrapper" suffix again.

https://codereview.chromium.org/25478003/diff/1/src/runtime.cc
File src/runtime.cc (right):

https://codereview.chromium.org/25478003/diff/1/src/runtime.cc#newcode1574
src/runtime.cc:1574: !isolate->MayNamedAccess(JSObject::cast(*obj),
We could use the handlified version here by using Handle<JSObject>::cast(obj)
and isolate->factory()->proto_string() instead.

https://codereview.chromium.org/25478003/diff/1/src/runtime.cc#newcode1742
src/runtime.cc:1742: obj->GetIsolate()->ReportFailedAccessCheck(*obj,
access_type);
nit: s/obj->GetIsolate()/isolate/

[Yang, 2013-10-01 15:48:29]

On 2013/10/01 14:46:18, Michael Starzinger wrote:
> LGTM with nits.
> 
> https://codereview.chromium.org/25478003/diff/1/src/isolate.h
> File src/isolate.h (right):
> 
> https://codereview.chromium.org/25478003/diff/1/src/isolate.h#newcode757
> src/isolate.h:757: bool MayNamedAccessWrapper(Handle<JSObject> receiver,
> At some point we want these functions to be the default implementations,
right?
> So I thing we should overload the name, otherwise we need to rename all
> call-sites to drop the "wrapper" suffix again.
> 
> https://codereview.chromium.org/25478003/diff/1/src/runtime.cc
> File src/runtime.cc (right):
> 
> https://codereview.chromium.org/25478003/diff/1/src/runtime.cc#newcode1574
> src/runtime.cc:1574: !isolate->MayNamedAccess(JSObject::cast(*obj),
> We could use the handlified version here by using Handle<JSObject>::cast(obj)
> and isolate->factory()->proto_string() instead.
> 
> https://codereview.chromium.org/25478003/diff/1/src/runtime.cc#newcode1742
> src/runtime.cc:1742: obj->GetIsolate()->ReportFailedAccessCheck(*obj,
> access_type);
> nit: s/obj->GetIsolate()/isolate/

I agree that overloading would be ideal. However, the function is being passed
to a helper function as a function pointer, so a compiler has issues with
overloaded function names. I haven't bothered with trying to find out how to do
it, so I added the wrapper-suffix as a temporary workaround.

[Yang, 2013-10-02 08:30:13]

Committed patchset #1 manually as r17062 (presubmit successful).