Name: rmT116609 Date: 05/14/2002


FULL PRODUCT VERSION :
java version "1.4.0"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.0-b92)
Java HotSpot(TM) Client VM (build 1.4.0-b92, mixed mode)

DESCRIPTION OF THE PROBLEM :
The new breakiterator which is currently giving warnings and is schedule for inclusion in the next JDK release is too aggressiove about recognizing HTMl tags as ending the first sentence. In particular, it thinks that <p> ends a paragraph
instead of starting one. This is incorrect, especially when using XHTML.

STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :

For example, consider this doccomment:

  /**
   * here's the first sentence
   * <p>
   */

Now suppose this comment is rewritten as proper XHTML. Then it would look as follows:

  /**
   * <p>
   * here's the first sentence
   * </p>
   */

The current javadoc tool works with this. However, with the breakiterator algorithm, it incorrectly thinks that the first <p> tag ends the first sentence instead of starting it.

The algorithm for recognizing the end of the first sentence should probably ignore possible first sentences that seem to contain only whitespace.

EXPECTED VERSUS ACTUAL BEHAVIOR :
With breakiterator, some first sentences are empty. Without them they're not.



---------- BEGIN SOURCE ----------
/**
 * <p>
 * Here's the first sentence.
 * </p>
 *
 * @author Elliotte Rusty Harold
 * @version $Revision: 1.0 $, $Date: 2002/05/13 19:17:10 $
 */
public class TestClass {

    
}
---------- END SOURCE ----------

(Review ID: 146533)
======================================================================