Keep maps when resetting elements.

Keep maps when resetting elements.

R=verwaest@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=21838

[Jakob Kummerow, 2014-06-12 16:36:22]

PTAL.

[Toon Verwaest, 2014-06-12 21:58:03]

lgtm with comment

https://codereview.chromium.org/335553004/diff/1/src/objects.cc
File src/objects.cc (right):

https://codereview.chromium.org/335553004/diff/1/src/objects.cc#newcode4861
src/objects.cc:4861: if (object->map() == heap->sloppy_arguments_elements_map())
{
Are you sure this can ever happen? We never handled that case before. It's
probably worthwhile to ask the O.o guys why they decided to add it. Might just
be defensive programming... I'd prefer to not have such code clutter the
code-base confusing us, if it's not necessary.

[Jakob Kummerow, 2014-06-13 11:56:18]

https://codereview.chromium.org/335553004/diff/1/src/objects.cc
File src/objects.cc (right):

https://codereview.chromium.org/335553004/diff/1/src/objects.cc#newcode4861
src/objects.cc:4861: if (object->map() == heap->sloppy_arguments_elements_map())
{
On 2014/06/12 21:58:03, Toon Verwaest wrote:
> Are you sure this can ever happen? We never handled that case before. It's
> probably worthwhile to ask the O.o guys why they decided to add it. Might just
> be defensive programming... I'd prefer to not have such code clutter the
> code-base confusing us, if it's not necessary.

Agreed, I don't see how this can happen, and it doesn't in the test suite.
Replaced it with a CHECK so clusterfuzz can correct us if we're wrong :-)

[Jakob Kummerow, 2014-06-13 12:12:50]

Committed patchset #2 manually as r21838 (presubmit successful).