Handlification of ArrayConstructorCommon().

Handlification of ArrayConstructorCommon().

R=yangguo@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=20024

[Igor Sheludko, 2014-03-17 15:29:27]

PTAL

[Yang, 2014-03-18 09:52:16]

LGTM with comments.

https://codereview.chromium.org/201303009/diff/20001/src/elements.cc
File src/elements.cc (right):

https://codereview.chromium.org/201303009/diff/20001/src/elements.cc#newcode1916
src/elements.cc:1916: Factory* factory = isolate->factory();
Small nit here. I don't think we need to fetch the factory upfront, given that
we don't actually use it in some of the code paths. In fact, factory is just a
reinterpret_cast of isolate.

https://codereview.chromium.org/201303009/diff/20001/src/objects.cc
File src/objects.cc (right):

https://codereview.chromium.org/201303009/diff/20001/src/objects.cc#newcode11353
src/objects.cc:11353: void JSArray::Initialize(Handle<JSArray> array, int
capacity, int length) {
Why is this still a "temporary wrapper"? Afaict the MaybeObject version of this
method is only used in test-heap.cc from within handlified code, why not replace
that last use and remove the MaybeObject version?

[Igor Sheludko, 2014-03-18 11:25:10]

https://codereview.chromium.org/201303009/diff/20001/src/elements.cc
File src/elements.cc (right):

https://codereview.chromium.org/201303009/diff/20001/src/elements.cc#newcode1916
src/elements.cc:1916: Factory* factory = isolate->factory();
On 2014/03/18 09:52:17, Yang wrote:
> Small nit here. I don't think we need to fetch the factory upfront, given that
> we don't actually use it in some of the code paths. In fact, factory is just a
> reinterpret_cast of isolate.

Done.

https://codereview.chromium.org/201303009/diff/20001/src/objects.cc
File src/objects.cc (right):

https://codereview.chromium.org/201303009/diff/20001/src/objects.cc#newcode11353
src/objects.cc:11353: void JSArray::Initialize(Handle<JSArray> array, int
capacity, int length) {
On 2014/03/18 09:52:17, Yang wrote:
> Why is this still a "temporary wrapper"? Afaict the MaybeObject version of
this
> method is only used in test-heap.cc from within handlified code, why not
replace
> that last use and remove the MaybeObject version?

Actually I did this in a next CL, but you are right. Done.

[Igor Sheludko, 2014-03-18 11:38:34]

Committed patchset #3 manually as r20024 (presubmit successful).