Mustang's API doc for java.text.Collator class (http://download.java.net/jdk6/docs/api/java/text/Collator.html )
includes a garbled character.

----
For example, in Czech, "e" and "f" are considered primary differences, while "e" and "?" are secondary differences, ----

'?' should be a 'e' with an accent.
The same character is correctly displayed in Java2 5.0's API doc.
(http://java.sun.com/j2se/1.5.0/docs/api/java/text/Collator.html)


I think this bug was caused during the generation process of API doc. The original document in Collator.java doesn't have any problem.
Please change the environment/locale which is used on the API doc generation process.
When I generated API doc for Collator using Mustang's javadoc tool myself, the character wasn't garbled. Therefore, I don't think this is a regression in javadoc tool nor Collator.java.