SUGGESTED FIX

Basically the change is
*** src/share/classes/java/util/regex/Pattern.java-     Wed Feb 14 13:55:32 2007
--- src/share/classes/java/util/regex/Pattern.java      Mon Apr 23 12:16:36 2007
*** 3764,3775 ****
--- 3764,3778 ----
          Node atom;
          Not(Node atom) {
              this.atom = atom;
          }
          boolean match(Matcher matcher, int i, CharSequence seq) {
+             if (i < matcher.to)
                  return !atom.match(matcher, i, seq)
                    && next.match(matcher, i+countChars(seq, i, 1), seq);
+           matcher.hitEnd = true;
+           return false;
          }
          boolean study(TreeInfo info) {
              info.minLength++;
              info.maxLength++;
              return next.study(info);
                                     

EVALUATION

This error arises, although other patterns are used.

Test case:
public final class Bug5088563{
    private static String REGEX;
    private static String INPUT;
    private static Pattern pattern;
    private static Matcher matcher;
    private static boolean found;

    public static void main(String[] argv) {
        initResources();
        processTest();
    }

    private static void initResources() {
       try {
           REGEX = "\\p{javaWhitespace}\\P{javaWhitespace}+\\p{javaWhitespace}";
           INPUT = "ASDASSAs    dsdssad fssdsASAd   sdsd sdsd  ddd";
       } catch (Exception ioe) {
             ioe.printStackTrace();
       }

        pattern = Pattern.compile(REGEX);
        matcher = pattern.matcher(INPUT);
        System.out.println("Current REGEX is: "+REGEX);
        System.out.println("Current INPUT is: "+INPUT);
    }

    private static void processTest() {
         try{
                while(matcher.find()) {
                    System.out.println("I found the text \"" + matcher.group() +
                                   "\" starting at index " + matcher.start() +
                                   " and ending at index " + matcher.end() + ".");
                    found = true;
                }
                if(!found)
                        System.out.println("No match found.");
                
             System.out.println("Test case passed");
             System.exit(0);
         }catch(Exception exe){
             exe.printStackTrace();
             System.out.println("Test case failed");      
             System.exit(1);
         }
    }
}

Output is:
Current REGEX is: \p{javaWhitespace}\P{javaWhitespace}+\p{javaWhitespace}
Current INPUT is: ASDASSAs    dsdssad fssdsASAd   sdsd sdsd  ddd
I found the text " dsdssad " starting at index 11 and ending at index 20.
I found the text " sdsd " starting at index 31 and ending at index 37.
java.lang.StringIndexOutOfBoundsException: String index out of range: 46
        at java.lang.String.charAt(String.java:558)
        at java.util.regex.Pattern.countChars(Pattern.java:2791)
        at java.util.regex.Pattern.access$000(Pattern.java:595)
        at java.util.regex.Pattern$Not.match(Pattern.java:3769)
        at java.util.regex.Pattern$Curly.match0(Pattern.java:4228)
        at java.util.regex.Pattern$Curly.match(Pattern.java:4202)
        at java.util.regex.Pattern$JavaTypeClass.match(Pattern.java:3600)
        at java.util.regex.Pattern$Start.match(Pattern.java:3019)
        at java.util.regex.Matcher.search(Matcher.java:1092)
        at java.util.regex.Matcher.find(Matcher.java:528)
        at Bug5088563.processTest(Bug5088563.java:59)
        at Bug5088563.main(Bug5088563.java:38)
Test case failed

Since initial bug was filed for tiger and got closed as not reproducible but was not fixed there, it needs to be reopen.
                                     

SUGGESTED FIX

boolean match(Matcher matcher, int i, CharSequence seq) {
            if (i < matcher.to)
                return !atom.match(matcher, i, seq)
		    && next.match(matcher, i+countChars(seq, i, 1), seq);
	    matcher.hitEnd = true;
	    return false;
        }
                                     

EVALUATION

This problem has been "accidently" fixed in Mustang by other regex rewrite work.
Root cause is the incorrect Not class implementation in 5.0 codebase, see the
suggested fix for a possible solution, if 5.0u fix is desired.
                                     

EVALUATION

Likely a documentation issue.

-- iag@sfbay 2004-09-18