This section is about regression tests of zope.testing itself. None of these tests make sense in Python 3.
>>> from zope.testing.doctest import *
LP #69988 and #144569 both assert that doctests fail when rendering non-ASCII output with a UnicodeDecodeError. However, this does not appear to be so:
>>> print u'abc'
abc
>>> print u'\xe9'.encode('utf-8')
é
Tests for LP #561568:
>>> v = u'foo\xe9bar'
>>> v
u'foo...bar'
>>> v.encode('utf-8')
'foo...bar'