Why don't check the range overflow when converting "Number" to "Float" or &-CodePudding

please look at the source code snippet that is from org.springframework.util.NumberUtils#convertNumberToTargetClass(), the method signature is <T extends Number> T convertNumberToTargetClass(number:Number, targetClass:Class<?>)

public static <T extends Number> T convertNumberToTargetClass(Number number, Class<T> targetClass)
            throws IllegalArgumentException {
    ... ...
    ... ...
    else if (Float.class == targetClass) {
        return (T) Float.valueOf(number.floatValue());
    }
    .... ....
    .... ....
}

When converting number to Float, it doesn't check whether the number overflows the range of Float or not. But it does check that when converting number to Byte, Integer or Short, as showing below:

    public static <T extends Number> T convertNumberToTargetClass(Number number, Class<T> targetClass)
            throws IllegalArgumentException {
        ... ...
        ... ...
        else if (Byte.class == targetClass) {
            long value = checkedLongValue(number, targetClass);
            if (value < Byte.MIN_VALUE || value > Byte.MAX_VALUE) {
                raiseOverflowException(number, targetClass);
            }
            return (T) Byte.valueOf(number.byteValue());
        }
        ... ...
        ... ...
    }

    private static long checkedLongValue(Number number, Class<? extends Number> targetClass) {
        BigInteger bigInt = null;
        if (number instanceof BigInteger) {
            bigInt = (BigInteger) number;
        }
        else if (number instanceof BigDecimal) {
            bigInt = ((BigDecimal) number).toBigInteger();
        }
        // Effectively analogous to JDK 8's BigInteger.longValueExact()
        if (bigInt != null && (bigInt.compareTo(LONG_MIN) < 0 || bigInt.compareTo(LONG_MAX) > 0)) {
            raiseOverflowException(number, targetClass);
        }
        return number.longValue();
    }

    private static void raiseOverflowException(Number number, Class<?> targetClass) {
        throw new IllegalArgumentException("Could not convert number ["   number   "] of type ["  
            number.getClass().getName()   "] to target class ["   targetClass.getName()   "]: overflow");
    }

I'm really curious why!

CodePudding user response：

Because there is no overflow.

If you do something like this:

Float aFloat = NumberUtils.convertNumberToTargetClass(Double.MAX_VALUE, Float.class);

The value of aFloat will be Infinity.

Checkout the API doc: https://docs.oracle.com/en/java/javase/17/docs/api/java.base/java/lang/Float.html