Not reporting minimal diff when words added about icdiff

[tomauty]

Looks like minimal differences like this are not reported by the underlying library being used, i.e. difflib._mdiff.

Per the mdiff spec:

def mdiff(self):
        r"""The difflib._mdiff() function returns an interator which returns a
        tuple: (from line tuple, to line tuple, boolean flag)

        from/to line tuple -- (line num, line text)
            line num -- integer or None (to indicate a context separation)
            line text -- original line text with following markers inserted:
                '\0+' -- marks start of added text
                '\0-' -- marks start of deleted text
                '\0^' -- marks start of changed text
                '\1' -- marks end of added/deleted/changed text

        boolean flag -- None indicates context separation, True indicates
            either "from" or "to" line contains a change, otherwise False.
        """
        return difflib._mdiff(self._get_old_text(), self._get_new_text())

And the output for your test files through mdiff is as follows:

(u'\x00-aaa\x01', u'\x00+aaa bbb ccc\x01')

So minimal differences would be a non-trivial change. Hope this helps shed some light on the issue. I also have a vested interest in this feature, but don't have enough python knowledge to make a change this substantial. May be good to look into a different difflib function?

[jeffkaufman]

Yes, this looks like a bug in difflib. Are you up for filing a bug against python?

[lurch]

Similar to my comment on #58, it seems that difflib is sometimes able to do the correct thing, depending on the input. No, I've got no idea why it works the way it does! ;-)

import difflib
>>> list(difflib.ndiff(['aaa'], ['aaa bbb ccc']))
['- aaa', '+ aaa bbb ccc']
>>> list(difflib.ndiff(['aaaa'], ['aaaa bbb ccc']))
['- aaaa', '+ aaaa bbb ccc']
>>> list(difflib.ndiff(['aaaaa'], ['aaaaa bbb ccc']))
['- aaaaa', '+ aaaaa bbb ccc']
>>> list(difflib.ndiff(['aaaaaa'], ['aaaaaa bbb ccc']))
['- aaaaaa', '+ aaaaaa bbb ccc']
>>> list(difflib.ndiff(['aaaaaaa'], ['aaaaaaa bbb ccc']))
['- aaaaaaa', '+ aaaaaaa bbb ccc']
>>> list(difflib.ndiff(['aaaaaaaa'], ['aaaaaaaa bbb ccc']))
['- aaaaaaaa', '+ aaaaaaaa bbb ccc']
>>> list(difflib.ndiff(['aaaaaaaaa'], ['aaaaaaaaa bbb ccc']))
['- aaaaaaaaa', '+ aaaaaaaaa bbb ccc']
>>> list(difflib.ndiff(['aaaaaaaaaa'], ['aaaaaaaaaa bbb ccc']))
['- aaaaaaaaaa', '+ aaaaaaaaaa bbb ccc']
>>> list(difflib.ndiff(['aaaaaaaaaaa'], ['aaaaaaaaaaa bbb ccc']))
['- aaaaaaaaaaa', '+ aaaaaaaaaaa bbb ccc']
>>> list(difflib.ndiff(['aaaaaaaaaaaa'], ['aaaaaaaaaaaa bbb ccc']))
['- aaaaaaaaaaaa', '+ aaaaaaaaaaaa bbb ccc', '?             ++++++++\n']
>>> list(difflib.ndiff(['aaaaaaaaaaaa'], ['aaaaaaaaaaaa bbbb ccc']))
['- aaaaaaaaaaaa', '+ aaaaaaaaaaaa bbbb ccc']
>>> list(difflib.ndiff(['aaaaaaaaaaaaa'], ['aaaaaaaaaaaaa bbbb ccc']))
['- aaaaaaaaaaaaa', '+ aaaaaaaaaaaaa bbbb ccc']
>>> list(difflib.ndiff(['aaaaaaaaaaaaaa'], ['aaaaaaaaaaaaaa bbbb ccc']))
['- aaaaaaaaaaaaaa', '+ aaaaaaaaaaaaaa bbbb ccc', '?               +++++++++\n']