Generalize map dependent codes array to allow multiple dependency groups.

Generalize map dependent codes array to allow multiple dependency groups.

This requires enabling of weak embedded maps in optimized code and is basis for the upcoming prototype check removal CL.

[ulan, 2013-02-08 13:51:39]

Please take a look.

[Michael Starzinger, 2013-02-14 10:40:52]

First round of comments. As discussed offline, could you rebase this CL so that
it contains Toon's comments from the related CL? I will then do the final round
ASAP.

https://codereview.chromium.org/12224035/diff/8001/src/objects-debug.cc
File src/objects-debug.cc (right):

https://codereview.chromium.org/12224035/diff/8001/src/objects-debug.cc#newco...
src/objects-debug.cc:608: DependentCodes::kWeaklyEmbeddedGroup, this));
Indentation is off.

https://codereview.chromium.org/12224035/diff/8001/src/objects.h
File src/objects.h (right):

https://codereview.chromium.org/12224035/diff/8001/src/objects.h#newcode4697
src/objects.h:4697: // the number of codes is less than the length of the array.
Can we also add a sentence stating that the order of the code objects within a
group is not preserved by this implementation?

https://codereview.chromium.org/12224035/diff/8001/src/objects.h#newcode4719
src/objects.h:4719: inline void clear_code_at(int i);
As discussed offline: Can we (visually) separate the low-level from the
high-level accessors? The low-level ones should only be used by the GC, adding a
comment in that regard would help to clarify that.

[ulan_google, 2013-02-14 11:59:06]

Thanks, I merge this CL to https://chromiumcodereview.appspot.com/12225099/ and
am going to close this issue.

https://codereview.chromium.org/12224035/diff/8001/src/objects.h
File src/objects.h (right):

https://codereview.chromium.org/12224035/diff/8001/src/objects.h#newcode4697
src/objects.h:4697: // the number of codes is less than the length of the array.
On 2013/02/14 10:40:52, Michael Starzinger wrote:
> Can we also add a sentence stating that the order of the code objects within a
> group is not preserved by this implementation?

Done.

https://codereview.chromium.org/12224035/diff/8001/src/objects.h#newcode4719
src/objects.h:4719: inline void clear_code_at(int i);
On 2013/02/14 10:40:52, Michael Starzinger wrote:
> As discussed offline: Can we (visually) separate the low-level from the
> high-level accessors? The low-level ones should only be used by the GC, adding
a
> comment in that regard would help to clarify that.

Done.

[Michael Starzinger, 2013-02-14 12:04:43]

Yep, that's fine, I'll continue my review there.