Implement correct OS and CC detection.

src/strtod.cc

 // Copyright 2012 the V8 project authors. All rights reserved.
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions are
 // met:
 //
 //     * Redistributions of source code must retain the above copyright
 //       notice, this list of conditions and the following disclaimer.
 //     * Redistributions in binary form must reproduce the above
 //       copyright notice, this list of conditions and the following
 //       disclaimer in the documentation and/or other materials provided
@@ skipping 158 matching lines @@
     int exponent = 0;
     *result = DiyFp(significand, exponent);
     *remaining_decimals = buffer.length() - read_digits;
   }
 }
 
 
 static bool DoubleStrtod(Vector<const char> trimmed,
                          int exponent,
                          double* result) {
-#if (V8_TARGET_ARCH_IA32 || defined(USE_SIMULATOR)) && !defined(_MSC_VER)
+#if (V8_TARGET_ARCH_IA32 || defined(USE_SIMULATOR)) && !V8_CC_MSVC
   // On x86 the floating-point stack can be 64 or 80 bits wide. If it is
   // 80 bits wide (as is the case on Linux) then double-rounding occurs and the
   // result is not accurate.
   // We know that Windows32 with MSVC, unlike with MinGW32, uses 64 bits and is
   // therefore accurate.
   // Note that the ARM and MIPS simulators are compiled for 32bits. They
   // therefore exhibit the same problem.
   return false;
 #endif
   if (trimmed.length() <= kMaxExactDoubleIntegerDecimalDigits) {
@@ skipping 243 matching lines @@
 
   double guess;
   if (DoubleStrtod(trimmed, exponent, &guess) ||
       DiyFpStrtod(trimmed, exponent, &guess)) {
     return guess;
   }
   return BignumStrtod(trimmed, exponent, guess);
 }
 
 } }  // namespace v8::internal