Which you can't do since both licensing and implementations disallow
including java.lang.* classes in your code.

That's what j2me gives you, so you have to live with it, unless you seek
out all the various unicode characters...

As a more complete answer, here's what GNU Classpath does:

 public static boolean isWhitespace(char ch)
 {
   int attr = readChar(ch);
   return ((((1 << (attr & TYPE_MASK))
             & ((1 << SPACE_SEPARATOR)
                | (1 << LINE_SEPARATOR)
                | (1 << PARAGRAPH_SEPARATOR))) != 0)
           && (attr & NO_BREAK_MASK) == 0)
     || (ch <= '\u001F' && ((1 << ch)
                            & ((1 << '\t')
                               | (1 << '\n')
                               | (1 << '\u000B')
                               | (1 << '\u000C')
                               | (1 << '\r')
                               | (1 << '\u001C')
                               | (1 << '\u001D')
                               | (1 << '\u001E')
                               | (1 << '\u001F'))) != 0);
 }

Where readChar is defined thusly:

 /**
  * Grabs an attribute offset from the Unicode attribute database. The
lower
  * 5 bits are the character type, the next 2 bits are flags, and the top
  * 9 bits are the offset into the attribute tables.
  *
  * @param ch the character to look up
  * @return the character's attribute offset and type
  * @see #TYPE_MASK
  * @see #NO_BREAK_MASK
  * @see #MIRROR_MASK
  * @see CharData#DATA
  * @see CharData#SHIFT
  */
 // Package visible for use in String.
 static char readChar(char ch)
 {
   // Perform 16-bit addition to find the correct entry in data.
   return data[(char) (blocks[ch >> CharData.SHIFT] + ch)];
 }

So I guess if you wanted to do all that in your j2me app, that would
make things work.  Actually, you'd have to do more - you'd have to copy
those constants yourself, because they are not defined in the j2me
character class.

Ciao,