Major Description
The method:
public void characters(String characters) throws SAXException
{ characters(characters.toCharArray(), 0, characters.length()); }does not check for null values.
On many code references a check is done "before" calling this methd. However on other sides, e.g. HSLFExtractor some values are not checked:
xhtml.characters( comment.getAuthor() );
which may be null.
The simplest fix would be to check for null on the handler and if it is null handle it as NOOP or insert the new UTF-8 "replacement char" to let the user decide.