Might be related to 2811

As I would like to use '...' in order not to care about some part of a
unicode string containing unicode literals (using ELLIPSIS).
The unicode string is inside the value of a dict returned by a function.

It is working on Python 2.4.4 and not working on Python 2.5.2.

I am attaching the output of doctest_problem.py on Python 2.5.2
in case it is necessary.

At the moment I haven't a clue where to look for a bug. Anybody? ;)

This has nothing to do with doctest. Starting with your script, I get:

>>> a = f()
>>> b = repr(a)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
UnicodeEncodeError: 'ascii' codec can't encode character u'\u0105' in
position 4: ordinal not in range(128)

__repr__() is supposed to return a str object. In your case, I suggest
to use %r instead of "%s":
    def __repr__(self):
        return '<%s %r>' % (self.__class__.__name__, self.x)