Package org.aspectj.org.eclipse.jdt.internal.compiler.util

Class FloatUtil

java.lang.Object

org.aspectj.org.eclipse.jdt.internal.compiler.util.FloatUtil

public class FloatUtil
extends Object

Internal utility for declaring with hexadecimal double and float literals.

Since:

3.1

Constructor Summary

Constructors

Constructor	Description
FloatUtil()

Method Summary

Modifier and Type	Method	Description
static double	valueOfHexDoubleLiteral(char[] source)	Returns the double value corresponding to the given hexadecimal floating-point double precision literal.
static float	valueOfHexFloatLiteral(char[] source)	Returns the float value corresponding to the given hexadecimal floating-point single precision literal.

Methods inherited from class java.lang.Object

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

Constructor Details

FloatUtil

public FloatUtil()

Method Details

valueOfHexFloatLiteral

public static float valueOfHexFloatLiteral(char[] source)

Returns the float value corresponding to the given hexadecimal floating-point single precision literal. The literal must be syntactically correct, and must be a float literal (end in a 'f' or 'F'). It must not include either leading or trailing whitespace or a sign.

This method returns the same answer as Float.parseFloat(new String(source)) does in JDK 1.5, except that this method returns Floal.NaN if it would underflow to 0 (parseFloat just returns 0). The method handles all the tricky cases, including fraction rounding to 24 bits and gradual underflow.

Parameters:

source - source string containing single precision hexadecimal floating-point literal

Returns:

the float value, including Float.POSITIVE_INFINITY if the non-zero value is too large to be represented, and Float.NaN if the non-zero value is too small to be represented

valueOfHexDoubleLiteral

public static double valueOfHexDoubleLiteral(char[] source)

Returns the double value corresponding to the given hexadecimal floating-point double precision literal. The literal must be syntactially correct, and must be a double literal (end in an optional 'd' or 'D'). It must not include either leading or trailing whitespace or a sign.

This method returns the same answer as Double.parseDouble(new String(source)) does in JDK 1.5, except that this method throw NumberFormatException in the case of overflow to infinity or underflow to 0. The method handles all the tricky cases, including fraction rounding to 53 bits and gradual underflow.

Parameters:

source - source string containing double precision hexadecimal floating-point literal

Returns:

the double value, including Double.POSITIVE_INFINITY if the non-zero value is too large to be represented, and Double.NaN if the non-zero value is too small to be represented