A helper function to speed up creation of V8 wrappers for DOM Nodes.

A helper function to speed up creation of V8 wrappers for DOM Nodes.


Committed: http://code.google.com/p/v8/source/detail?r=2260

[Mads Ager (chromium), 2009-06-23 09:23:43]

http://codereview.chromium.org/141044/diff/1002/4
File include/v8.h (right):

http://codereview.chromium.org/141044/diff/1002/4#newcode1179
Line 1179: bool CloneElementAt(uint32_t index, Handle<Object>* result);
Please add a comment that explains this method including the fact that it can
fail.

I would remove the bool return value and just have the signature:

Handle<Object> CloneElementAt(uint32_t index);

Returning an empty handle will then signal that the cloning failed.

http://codereview.chromium.org/141044/diff/1002/5
File src/api.cc (right):

http://codereview.chromium.org/141044/diff/1002/5#newcode3026
Line 3026: i::Object* cloned =
i::Heap::CopyJSObject(i::JSObject::cast(paragon));
Why not use the handle based version i::Copy instead and let that handle
allocation failures?  You can model this after the Object::Clone method.

[antonm, 2009-06-23 13:56:23]

Thanks a lot, Mads!

I've updated the companion CL as well.

http://codereview.chromium.org/141044/diff/1002/4
File include/v8.h (right):

http://codereview.chromium.org/141044/diff/1002/4#newcode1179
Line 1179: bool CloneElementAt(uint32_t index, Handle<Object>* result);
On 2009/06/23 09:23:43, Mads Ager wrote:
> Please add a comment that explains this method including the fact that it can
> fail.
> 
> I would remove the bool return value and just have the signature:
> 
> Handle<Object> CloneElementAt(uint32_t index);
> 
> Returning an empty handle will then signal that the cloning failed.

Done.

http://codereview.chromium.org/141044/diff/1002/5
File src/api.cc (right):

http://codereview.chromium.org/141044/diff/1002/5#newcode3026
Line 3026: i::Object* cloned =
i::Heap::CopyJSObject(i::JSObject::cast(paragon));
On 2009/06/23 09:23:43, Mads Ager wrote:
> Why not use the handle based version i::Copy instead and let that handle
> allocation failures?  You can model this after the Object::Clone method.

I started with it actually.  Heap version runs somewhat (not dramatically
faster).  Anyway, reverted to Copy

Done.

[Mads Ager (chromium), 2009-06-24 08:42:53]

LGTM