Make v8 compile on VS2013.

Make v8 compile on VS2013.

VS2013 contains a number of improvements, most notably the addition of all C99 math functions.

I'm a little bit concerned about the change I had to make in cpu-profiler.cc, but I spent quite a bit of time looking at it and was unable to figure out any rational explanation for the warning.  It's possible it's spurious.  Since it seems like a useful warning in general though, I chose not to disable globally at the gyp level.

I do think someone with expertise here should probably try to determine if this is a legitimate warning.

BUG=288948
R=dslomov@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=16775

[zturner, 2013-09-13 23:31:38]

On 2013/09/13 23:14:07, zturner wrote:

Not sure why the trybot is unable to apply the patch.  Am I doing something
wrong?  The Base URL and relative paths look ok to me.

[Dmitry Lomov (no reviews), 2013-09-14 19:32:07]

On 2013/09/13 23:31:38, zturner wrote:
> On 2013/09/13 23:14:07, zturner wrote:
> 
> Not sure why the trybot is unable to apply the patch.  Am I doing something
> wrong?  The Base URL and relative paths look ok to me.

trybots do not work for v8 (your best bet is to build and run all tests locally)

[zturner, 2013-09-16 16:44:02]

On 2013/09/14 19:32:07, Dmitry Lomov (chromium) wrote:
> On 2013/09/13 23:31:38, zturner wrote:
> > On 2013/09/13 23:14:07, zturner wrote:
> > 
> > Not sure why the trybot is unable to apply the patch.  Am I doing something
> > wrong?  The Base URL and relative paths look ok to me.
> 
> trybots do not work for v8 (your best bet is to build and run all tests
locally)

Well really running the trybot was just a formality.  The only test was to make
sure it compiled on the trybots, since there's no functional change involved.

[Dmitry Lomov (no reviews), 2013-09-17 09:54:36]

lgtm (sorry for late reply)
I'll land for you

[Dmitry Lomov (no reviews), 2013-09-17 15:26:46]

Committed patchset #1 manually as r16775 (presubmit successful).

[Sven Panne, 2014-01-09 07:36:15]

Actually, the cpu-profiler.cc part is wrong. :-/ SamplingCircularQueue, a member
of ProfilerEventsProcessor, actually has a 64-byte alignment requirement, so
VS2013 is correct. We have similar problems in test-cpu-profiler.cc, and all
those problems are platform-independent, because no platform guarantees such an
alignment for new/malloc. Probably we're just lucky because
ProfilerEventsProcessor is so big that the underlying memory management system
allocates new pages from the OS, anyway, having the required alignment.

I'll come up with a real fix...

[Dmitry Lomov (no reviews), 2014-01-10 08:33:32]

On 2014/01/09 07:36:15, Sven Panne wrote:
> Actually, the cpu-profiler.cc part is wrong. :-/ SamplingCircularQueue, a
member
> of ProfilerEventsProcessor, actually has a 64-byte alignment requirement, so
> VS2013 is correct. We have similar problems in test-cpu-profiler.cc, and all
> those problems are platform-independent, because no platform guarantees such
an
> alignment for new/malloc. Probably we're just lucky because
> ProfilerEventsProcessor is so big that the underlying memory management system
> allocates new pages from the OS, anyway, having the required alignment.
> 
> I'll come up with a real fix...

To be fair, the problem with cpu-profiler.cc existed before this CL, it is just
that VS2013 is kind enough to warn us that alignment is wrong. 
Thanks for taking a deep look!