CJKCodecs is distributed with utf-7 and utf-8 codec in spite of Python
has their own already. Here're my cowardly rationales for that.

 - Python UTF-7 codec can't encode and/or decode surrogate pair.

 - Python UTF-7 codec can't decode long shifted sequence.

 - Python UTF-7 codec isn't stateful, so its StreamReader and
   StreamWriter can't work correctly.

 - Python UTF-8 codec is slightly broken for StreamReader's readline
   and readlines method calls. For example,

    >>> import StringIO, codecs
    >>> c = codecs.getreader('utf-8')(StringIO.StringIO("Python\xed\x8c\x8c\xec\x9d\xb4\xec\x8d\xac"))
    >>> c.readline(1), c.readline(1), c.readline(1)
    u'P', u'y', u't'
    >>> c.readline(1), c.readline(1), c.readline(1)
    (u'h', u'o', u'n')
    >>> c.readline(1), c.readline(1), c.readline(1)
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
      File "/usr/local/lib/python2.2/codecs.py", line 252, in readline
        return self.decode(line, self.errors)[0]
    UnicodeError: UTF-8 decoding error: unexpected end of data