[PATCH] ldb: The test api.py should not rely on order of entries in

[Lukas Slebodnik]

On (28/01/19 14:20), Douglas Bagnall via samba-technical wrote:
>On 23/01/19 3:00 AM, Lukas Slebodnik wrote:
>
>> I've just realized that unreliable order of entries is already solved
>> in the test for python3
>> 
>>     def test_repr(self):
>>         self.msg.dn = ldb.Dn(ldb.Ldb(), "dc=foo29")
>>         self.msg["dc"] = b"foo"
>>         if PY3:
>>             self.assertIn(repr(self.msg), [
>>                 "Message({'dn': Dn('dc=foo29'), 'dc': MessageElement([b'foo'])})",
>>                 "Message({'dc': MessageElement([b'foo']), 'dn': Dn('dc=foo29')})",
>>             ])
>>             self.assertIn(repr(self.msg.text), [
>>                 "Message({'dn': Dn('dc=foo29'), 'dc': MessageElement([b'foo'])}).text",
>>                 "Message({'dc': MessageElement([b'foo']), 'dn': Dn('dc=foo29')}).text",
>>             ])
>> 
>> Therefore we can use the same approach for python2. There will be just
>> difference between bytes and strings.
>
>Yes. RB+.
>
>Actually with 3.6+, the order of entries should be reliable given the
>new python dict behaviour (the dn always comes first).
>

It worked just by a chance in 3.6 and in 3.7 it is officially supported :-)
https://docs.python.org/3/whatsnew/3.7.html
"""
  the insertion-order preservation nature of dict objects has been declared to be
  an official part of the Python language spec.
"""

Thank you very much for review.

LS