A DESCRIPTION OF THE PROBLEM :
https://docs.oracle.com/en/java/javase/17/docs/api/java.base/java/util/StringTokenizer.html#%3Cinit%3E(java.lang.String,java.lang.String,boolean) said: "Each delimiter is returned as a string of length one." This is not correct if any of the delimiter is a valid Unicode surrogate pair since the returned string will be of length two because the delimiter is represented by two code units.

EXPECTED VERSUS ACTUAL BEHAVIOR :
EXPECTED -
"Each delimiter is returned as a string of the code unit(s) of the delimiter."

Or remove "Each delimiter is returned as a string of length one." and clarify that "characters" in StringTokenizer documentation context refers to Unicode code points like other documentation, e.g., that of String: "The String class provides methods for dealing with Unicode code points (i.e., characters), in addition to those for dealing with Unicode code units (i.e., char values)." - https://docs.oracle.com/en/java/javase/17/docs/api/java.base/java/lang/String.html.
ACTUAL -
"Each delimiter is returned as a string of length one."

---------- BEGIN SOURCE ----------
import java.util.StringTokenizer;

public class StringTokenizerPlayground {

  public static void main(String[] args) {
    final var s = "\uD83D\uDE00"; // Grinning Face
    final var tokenizer = new StringTokenizer(s, s, true);

    final var tokenCount = tokenizer.countTokens();

    if (tokenCount != 1) {
      throw new AssertionError();
    }

    final var token = tokenizer.nextToken();

    if (token.length() != 2) {
      throw new AssertionError();
    }

    if (!token.equals(s)) {
      throw new AssertionError();
    }
  }
}
---------- END SOURCE ----------

FREQUENCY : always