Replace math.h with cmath

src/win32-math.cc

Index: src/win32-math.cc
diff --git a/src/win32-math.cc b/src/win32-math.cc
index 3410872bb5a4554e83ca56b1a86612aa4fabe608..9ffc4ea73bd77a4c44d0148c25ae8421610aea5f 100644
--- a/src/win32-math.cc
+++ b/src/win32-math.cc
@@ -35,21 +35,14 @@
 #define V8_WIN32_HEADERS_FULL
 #include "win32-headers.h"
 #include <limits.h>        // Required for INT_MAX etc.
-#include <math.h>
 #include <float.h>         // Required for DBL_MAX and on Win32 for finite()
+#include <cmath>
 #include "win32-math.h"
 
 #include "checks.h"
 
-namespace v8 {
-
-// Test for finite value - usually defined in math.h
-int isfinite(double x) {
-  return _finite(x);
-}
-
-}  // namespace v8
 
+namespace std {
 
 // Test for a NaN (not a number) value - usually defined in math.h
 int isnan(double x) {
@@ -63,6 +56,12 @@ int isinf(double x) {
 }
 
 
+// Test for finite value - usually defined in math.h
+int isfinite(double x) {
+  return _finite(x);
+}
+
+
 // Test if x is less than y and both nominal - usually defined in math.h
 int isless(double x, double y) {
   return isnan(x) || isnan(y) ? 0 : x < y;
@@ -103,4 +102,6 @@ int signbit(double x) {
     return x < 0;
 }
 
+}  // namespace std
+
 #endif  // _MSC_VER