Package com.google.javascript.jscomp.base

Class JSCompDoubles

  • public final class JSCompDoubles
extends java.lang.Object
    Basic double functions used by JSComp.

    • Method Summary

      All Methods Static Methods Concrete Methods 
      Modifier and Type Method Description
      static int ecmascriptToInt32​(double number)
      The ECMAScript ToInt32 abstract operation.
      static int ecmascriptToUint32​(double number)
      The ECMAScript ToUint32 abstract operation.
      static boolean isAtLeastIntegerPrecision​(double x)
      Does `x` have precision down to the "ones column"?
      static boolean isEitherZero​(double x)
      Is `x` positive or negative zero?
      static boolean isExactInt32​(double x)
      Can a 32 bit int exactly represent `x`?
      static boolean isExactInt64​(double x)
      Can a 64 bit int exactly represent `x`?
      static boolean isMathematicalInteger​(double x)
      Does `x` exactly represent an value with no fractional part?
      static boolean isNegative​(double x)
      Does `x` carry a negative sign?
      static boolean isPositive​(double x)
      Does `x` not carry a negative sign?

      • Methods inherited from class java.lang.Object

        clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

    • Method Detail

      • isExactInt32

        public static boolean isExactInt32​(double x)
        Can a 32 bit int exactly represent `x`?

        Many double values are not exact integers. Many that are integers are too large to fit into a Java int.

        This function does not guarantee that a value can be round-tripped from double to int to double and have an identical bit pattern. Notably, 0.0 and -0.0 both represent exactly 0.

      • isExactInt64

        public static boolean isExactInt64​(double x)
        Can a 64 bit int exactly represent `x`?

        Many double values are not exact integers. Many that are integers are too large to fit into a Java long.

        This function does not guarantee that a value can be round-tripped from double to long to double and have an identical bit pattern. Notably, 0.0 and -0.0 both represent exactly 0.

      • isMathematicalInteger

        public static boolean isMathematicalInteger​(double x)
        Does `x` exactly represent an value with no fractional part?

        The value may be too large to fit in a primitive integral type, such as long.

        Returns false for NaN and Infinity.

        This should behave identically to Guava DoubleMath.isMathematicalInteger but is J2CL compatible.

      • isAtLeastIntegerPrecision

        public static boolean isAtLeastIntegerPrecision​(double x)
        Does `x` have precision down to the "ones column"?

        A double can hold exact integer values that are very large, but to do so it may loose precision at the scale of "ones". That is "largeDouble + 1.0 == largeDouble" may be true.

        Returns false for NaN and Infinity.

      • isNegative

        public static boolean isNegative​(double x)
        Does `x` carry a negative sign?

        Because -0.0 == 0.0, it is not enough to check `x < 0.0` to determine if a double is negative. This function identifies the -0.0 case.

      • isPositive

        public static boolean isPositive​(double x)
        Does `x` not carry a negative sign?

        Because -0.0 == 0.0, it is not enough to check `x < 0.0` to determine if a double is negative. This function identifies the -0.0 case.

      • isEitherZero

        public static boolean isEitherZero​(double x)
        Is `x` positive or negative zero?

      • ecmascriptToInt32

        public static int ecmascriptToInt32​(double number)
        The ECMAScript ToInt32 abstract operation.

        See https://262.ecma-international.org/5.1/#sec-9.5

      • ecmascriptToUint32

        public static int ecmascriptToUint32​(double number)
        The ECMAScript ToUint32 abstract operation.

        Java has no uint types, so the caller must remember to treat the returned bits as a uint.

        See https://262.ecma-international.org/5.1/#sec-9.6