Re-enable freezing of Object.observe changeRecords now that freezing is faster.

Re-enable freezing of Object.observe changeRecords now that freezing is faster.

Committed: http://code.google.com/p/v8/source/detail?r=14848

[adamk, 2013-05-23 21:12:39]

https://codereview.chromium.org/15898004/diff/1/src/object-observe.js
File src/object-observe.js (right):

https://codereview.chromium.org/15898004/diff/1/src/object-observe.js#newcode312
src/object-observe.js:312: ObjectFreeze(changeRecord.removed);
I'd leave this out for now, as it basically doubles the cost of creating a
splice record (create the removed array, then create a dictionary backing store
and copy everything over).

[rafaelw, 2013-05-23 22:53:18]

https://codereview.chromium.org/15898004/diff/1/src/object-observe.js
File src/object-observe.js (right):

https://codereview.chromium.org/15898004/diff/1/src/object-observe.js#newcode312
src/object-observe.js:312: ObjectFreeze(changeRecord.removed);
I went back and tried as hard as I could to create a case where freezing the
removed array impacted the over-all time noticeably -- and failed to produce
anything that was obviously not in the noise.

I think it's fine to leave this in.

On 2013/05/23 21:12:39, adamk wrote:
> I'd leave this out for now, as it basically doubles the cost of creating a
> splice record (create the removed array, then create a dictionary backing
store
> and copy everything over).

[adamk, 2013-05-23 22:55:49]

lgtm

https://codereview.chromium.org/15898004/diff/1/src/object-observe.js
File src/object-observe.js (right):

https://codereview.chromium.org/15898004/diff/1/src/object-observe.js#newcode312
src/object-observe.js:312: ObjectFreeze(changeRecord.removed);
On 2013/05/23 22:53:18, rafaelw wrote:
> I went back and tried as hard as I could to create a case where freezing the
> removed array impacted the over-all time noticeably -- and failed to produce
> anything that was obviously not in the noise.
> 
> I think it's fine to leave this in.

Sounds good to me, thanks for doing the footwork to verify this.

> On 2013/05/23 21:12:39, adamk wrote:
> > I'd leave this out for now, as it basically doubles the cost of creating a
> > splice record (create the removed array, then create a dictionary backing
> store
> > and copy everything over).
>

[rossberg, 2013-05-27 11:40:46]

lgtm

[rossberg, 2013-05-28 11:34:37]

Committed patchset #1 manually as r14848 (presubmit successful).