Fix passing required=True to DjangoConnectionField

[alexkirsz]

Passing required=True to an instance of DjangoConnectionField currently raises the following exception:

  ...
  File "/env/lib/python3.6/site-packages/graphql/type/typemap.py", line 106, in reducer
    field_map = type.fields
  File "/env/lib/python3.6/site-packages/graphql/pyutils/cached_property.py", line 22, in __get__
    value = obj.__dict__[self.func.__name__] = self.func(obj)
  File "/env/lib/python3.6/site-packages/graphql/type/definition.py", line 221, in fields
    return define_field_map(self, self._fields)
  File "/env/lib/python3.6/site-packages/graphql/type/definition.py", line 235, in define_field_map
    field_map = field_map()
  File "/env/lib/python3.6/site-packages/graphene/types/typemap.py", line 274, in construct_fields_for_type
    map = self.reducer(map, field.type)
  File "/env/lib/python3.6/site-packages/graphene-django/graphene_django/fields.py", line 61, in type
    _type, DjangoObjectType
TypeError: issubclass() arg 1 must be a class

The reason for the exception is that DjangoConnectionField does not expect its type to be wrapped in a NonNull instance. This PR ensures that it handles that case properly.

[coveralls]

Coverage decreased (-0.3%) to 94.071% when pulling 8beadc7 on acu:fix-connection-field-required into ea2cd98 on graphql-python:master.

[coveralls]

Coverage decreased (-0.3%) to 94.071% when pulling 8beadc7 on acu:fix-connection-field-required into ea2cd98 on graphql-python:master.

[coveralls]

Coverage decreased (-0.3%) to 94.071% when pulling 8beadc7 on acu:fix-connection-field-required into ea2cd98 on graphql-python:master.

[coveralls]

Coverage decreased (-0.3%) to 94.071% when pulling 8beadc7 on acu:fix-connection-field-required into ea2cd98 on graphql-python:master.

[coveralls]

Coverage decreased (-0.4%) to 93.934% when pulling b49d315 on acu:fix-connection-field-required into ea2cd98 on graphql-python:master.

[phalt]

8 spaces should be 4!

[alexkirsz]

@phalt Done!

[phalt]

This looks good to me - @alexkirsz this isn't a breaking change is it?

[alexkirsz]

@phalt The previous behaviour was to raise an exception, so I don't believe there's any way in which this could break non-already broken code :)

[phalt]

@alexkirsz approval from me then!