Merge V8 at 3.9.24.13
Bug: 5688872
Change-Id: Id0aa8d23375030494d3189c31774059c0f5398fc
diff --git a/src/conversions.h b/src/conversions.h
index e51ad65..70559c9 100644
--- a/src/conversions.h
+++ b/src/conversions.h
@@ -28,8 +28,6 @@
#ifndef V8_CONVERSIONS_H_
#define V8_CONVERSIONS_H_
-#include <limits>
-
#include "utils.h"
namespace v8 {
@@ -47,14 +45,14 @@
const int kMaxSignificantDigits = 772;
-static inline bool isDigit(int x, int radix) {
+inline bool isDigit(int x, int radix) {
return (x >= '0' && x <= '9' && x < '0' + radix)
|| (radix > 10 && x >= 'a' && x < 'a' + radix - 10)
|| (radix > 10 && x >= 'A' && x < 'A' + radix - 10);
}
-static inline double SignedZero(bool negative) {
+inline double SignedZero(bool negative) {
return negative ? -0.0 : 0.0;
}
@@ -63,16 +61,16 @@
// rounding towards zero.
// The result is unspecified if x is infinite or NaN, or if the rounded
// integer value is outside the range of type int.
-static inline int FastD2I(double x) {
+inline int FastD2I(double x) {
// The static_cast convertion from double to int used to be slow, but
// as new benchmarks show, now it is much faster than lrint().
return static_cast<int>(x);
}
-static inline unsigned int FastD2UI(double x);
+inline unsigned int FastD2UI(double x);
-static inline double FastI2D(int x) {
+inline double FastI2D(int x) {
// There is no rounding involved in converting an integer to a
// double, so this code should compile to a few instructions without
// any FPU pipeline stalls.
@@ -80,7 +78,7 @@
}
-static inline double FastUI2D(unsigned x) {
+inline double FastUI2D(unsigned x) {
// There is no rounding involved in converting an unsigned integer to a
// double, so this code should compile to a few instructions without
// any FPU pipeline stalls.
@@ -89,15 +87,15 @@
// This function should match the exact semantics of ECMA-262 9.4.
-static inline double DoubleToInteger(double x);
+inline double DoubleToInteger(double x);
// This function should match the exact semantics of ECMA-262 9.5.
-static inline int32_t DoubleToInt32(double x);
+inline int32_t DoubleToInt32(double x);
// This function should match the exact semantics of ECMA-262 9.6.
-static inline uint32_t DoubleToUint32(double x) {
+inline uint32_t DoubleToUint32(double x) {
return static_cast<uint32_t>(DoubleToInt32(x));
}