Name: ooR10006 Date: 03/19/2001


Although the API spec for class java.lang.Character was correctly changed in jdk1.4.0beta-b56,
(see 4404977) the following 3 methods return wrong results according to the changed spec:

1. The spec for Character.isWhitespace(ch) says:
"Determines if the specified character is white space according to Java. A character is
a Java whitespace character if and only if it satisfies one of the following criteria:
   It is a Unicode space character (SPACE_SEPARATOR, LINE_SEPARATOR, or
   PARAGRAPH_SEPARATOR) but is not also a non-breaking space (\u00A0, \u2007, \u202F)..."

However, method isWhitespace(char) returns true for 0x2007, 0x202F.
the following test shows this:
 
class Test {
    public static void main(String[] args){
        System.out.println(Character.isWhitespace((char)0x2007) + " " +
        Character.isWhitespace((char)0x202F));
        return;
    }
}
% jdk1.4.0beta-b56/solsparc/bin/java Test
true true

2. method getDirectionality(ch) returns value 0 for all undefined in UnicodeData characters
instead of DIRECTIONALITY_UNDEFINED i.e. -1.
The following example shows this:
class Test {
    public static void main(String[] args){
        for (int i = 0; i < 0xFFFF; ++i){
            if (!Character.isDefined((char)i)){
                System.out.println(Character.getDirectionality((char)i));
            }
        }
        return;
    }
}

The bug 4404977 which describes all the changes of Character spec has state 'integrated',
the jdk1.4.0beta-b56 API spec is indeed changed, but the implementation doesn't conform
to the spec.

Due to this new Merlin JCK tests fail:
api/java_lang/Character/index.html#getAttrFullRange[Character2086]
api/java_lang/Character/index.html#attributesFullRange[Character2070]
api/java_lang/Character/index.html#attributesFullRange[Character2079]
 
======================================================================