Replace our home-grown BitCast with bit_cast from Chrome/Google3.

src/x87/lithium-codegen-x87.cc

Index: src/x87/lithium-codegen-x87.cc
diff --git a/src/x87/lithium-codegen-x87.cc b/src/x87/lithium-codegen-x87.cc
index 90e9c90787d5f4fb72a6028ef74512521d9d2a3a..6a3f991fbc4d1745126062f29c4064af9480e5ed 100644
--- a/src/x87/lithium-codegen-x87.cc
+++ b/src/x87/lithium-codegen-x87.cc
@@ -1908,7 +1908,7 @@ void LCodeGen::DoConstantS(LConstantS* instr) {
 
 void LCodeGen::DoConstantD(LConstantD* instr) {
   double v = instr->value();
-  uint64_t int_val = BitCast<uint64_t, double>(v);
+  uint64_t int_val = bit_cast<uint64_t, double>(v);
   int32_t lower = static_cast<int32_t>(int_val);
   int32_t upper = static_cast<int32_t>(int_val >> (kBitsPerInt));
   DCHECK(instr->result()->IsDoubleRegister());
@@ -4088,7 +4088,7 @@ void LCodeGen::DoStoreKeyedFixedDoubleArray(LStoreKeyed* instr) {
     // This means we should store the (double) hole. No floating point
     // registers required.
     double nan_double = FixedDoubleArray::hole_nan_as_double();
-    uint64_t int_val = BitCast<uint64_t, double>(nan_double);
+    uint64_t int_val = bit_cast<uint64_t, double>(nan_double);
     int32_t lower = static_cast<int32_t>(int_val);
     int32_t upper = static_cast<int32_t>(int_val >> (kBitsPerInt));