It turned out the Apache Commons Lang project, not Sun, provides the decoder and the encoder. I remembered its value from the time I was looking for an XML encoder to use in my JSP page.
http://commons.apache.org/lang/api/o
http://svn.apache.org/viewvc/commons/pr
I started my search when I realized that a brute force approach such as
s.replace("\\\\", "\\").replace("\\n", "\n") would fail to correctly decode a legitimate sequence of 3 characters '\\', '\\' and 'n' into a string of 2 characters '\\' and 'n'. This is because multiple .replace() invocations apply per-character decoding more than once, which is wrong.