Stack sampling profiler to profile on Mac. TEST-- not for commit.

Stack sampling profiler to profile on Mac. TEST-- not for commit.

BUG=

[sydli, 2015-09-11 23:51:31]

Patchset #2 (id:20001) has been deleted

[Robert Sesek, 2015-09-14 18:55:53]

rsesek@chromium.org changed reviewers:
+ rsesek@chromium.org

[Robert Sesek, 2015-09-14 18:55:54]

Just some high-level comments.

https://codereview.chromium.org/1342453003/diff/100001/base/profiler/native_s...
File base/profiler/native_stack_sampler_posix.cc (right):

https://codereview.chromium.org/1342453003/diff/100001/base/profiler/native_s...
base/profiler/native_stack_sampler_posix.cc:69: void
ConvertIdentifierToString(const unsigned char identifier[16],
Can you use HexEncode() from base/strings/string_number_conversions.h instead ?

https://codereview.chromium.org/1342453003/diff/100001/base/profiler/native_s...
base/profiler/native_stack_sampler_posix.cc:240: GetUniqueId(inf.dli_fname,
&unique_id);
Touching disk here is going to be a huge hit on some machines and under some
circumstances. But you also do not need to. Use dli_fbase instead, which is a
pointer to the mach_header. The image, and all of its information, is already
loaded into memory. There's no reason to access disk for it.

https://codereview.chromium.org/1342453003/diff/100001/base/profiler/native_s...
base/profiler/native_stack_sampler_posix.cc:266: class NativeStackSamplerMac :
public NativeStackSampler {
This should move into native_stack_sampler_mac.cc (and the impl above), since
it's not POSIX.

https://codereview.chromium.org/1342453003/diff/100001/base/profiler/native_s...
base/profiler/native_stack_sampler_posix.cc:288: // (See
base/threading/platform_thread_posix.cc:128)
This seems like a bug. May want to file one for this to redefine it as a
mach_port_t, since that's what it really is.