Enable incremental marking when --expose-gc is turned on.

Enable incremental marking when --expose-gc is turned on.

BUG=
R=mstarzinger@chromium.org, ulan@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=18691

[ulan, 2014-01-20 14:04:45]

the idle notification changes lgtm

[Michael Starzinger, 2014-01-20 14:09:30]

LGTM with one comment.

https://codereview.chromium.org/139343007/diff/1/src/heap.cc
File src/heap.cc (right):

https://codereview.chromium.org/139343007/diff/1/src/heap.cc#newcode5609
src/heap.cc:5609: kReduceMemoryFootprintMask | kAbortIncrementalMarkingMask,
This not only removes the special handling of --expose-gc but also changes
semantics of IdleNotification. I have no beef with this change in particular,
but maybe we should do that in a separate CL, so that --noexpose-gc semantics
stay unchanged with this CL.

[Hannes Payer (out of office), 2014-01-20 14:14:14]

Committed patchset #3 manually as r18691 (presubmit successful).

[Hannes Payer (out of office), 2014-01-20 14:28:42]

https://codereview.chromium.org/139343007/diff/1/src/heap.cc
File src/heap.cc (right):

https://codereview.chromium.org/139343007/diff/1/src/heap.cc#newcode5609
src/heap.cc:5609: kReduceMemoryFootprintMask | kAbortIncrementalMarkingMask,
On 2014/01/20 14:09:30, Michael Starzinger wrote:
> This not only removes the special handling of --expose-gc but also changes
> semantics of IdleNotification. I have no beef with this change in particular,
> but maybe we should do that in a separate CL, so that --noexpose-gc semantics
> stay unchanged with this CL.

Done, makes sense. I will do that.