Implement the changes proposed in patch #413333. unicode(obj) now
works just like str(obj) in that it tries __str__/tp_str on the object
in case it finds that the object is not a string or buffer.
2 files changed