Use std:: on symbols declared in C++-style C headers.

src/assembler.cc

Index: src/assembler.cc
diff --git a/src/assembler.cc b/src/assembler.cc
index 4ac2fb49eafc41429f4deaf91d16573126c296cb..b619b6cc6be486c40142d816ce9b4f10983ad887 100644
--- a/src/assembler.cc
+++ b/src/assembler.cc
@@ -938,7 +938,7 @@ void ExternalReference::InitializeMathExpData() {
     // The rest is black magic. Do not attempt to understand it. It is
     // loosely based on the "expd" function published at:
     // http://herumi.blogspot.com/2011/08/fast-double-precision-exponential.html
-    const double constant3 = (1 << kTableSizeBits) / log(2.0);
+    const double constant3 = (1 << kTableSizeBits) / std::log(2.0);
     math_exp_constants_array[3] = constant3;
     math_exp_constants_array[4] =
         static_cast<double>(static_cast<int64_t>(3) << 51);
@@ -949,7 +949,7 @@ void ExternalReference::InitializeMathExpData() {
 
     math_exp_log_table_array = new double[kTableSize];
     for (int i = 0; i < kTableSize; i++) {
-      double value = pow(2, i / kTableSizeDouble);
+      double value = std::pow(2, i / kTableSizeDouble);
       uint64_t bits = BitCast<uint64_t, double>(value);
       bits &= (static_cast<uint64_t>(1) << 52) - 1;
       double mantissa = BitCast<double, uint64_t>(bits);
@@ -1392,7 +1392,7 @@ ExternalReference ExternalReference::math_log_double_function(
     Isolate* isolate) {
   typedef double (*d2d)(double x);
   return ExternalReference(Redirect(isolate,
-                                    FUNCTION_ADDR(static_cast<d2d>(log)),
+                                    FUNCTION_ADDR(static_cast<d2d>(std::log)),
                                     BUILTIN_FP_CALL));
 }
 
@@ -1477,7 +1477,7 @@ double power_double_double(double x, double y) {
   if (std::isnan(y) || ((x == 1 || x == -1) && std::isinf(y))) {
     return OS::nan_value();
   }
-  return pow(x, y);
+  return std::pow(x, y);
 }