I had some UTF8 problems. It turned out that the DbaseFileReader.java attached by limonolith just fixed all my problems.
My attached utf8supportAndTests.tar.gz contains
1. a small and royalty-free shapefile with an UTF8 encoded DBF
2. a patch that
** adds some JUnit-Tests that access the utf8 attributes
** replaces the existing DbaseFileReader with the one given by limonolith
** adds the optimizations from http://jira.codehaus.org/browse/GEOT-2683
to the new DbfFileReader.
"mvn test" just passed... so i would be happy to see this comitted next week.