Behaviors are different when data is a list(dict) and text

When data is unicode, it gives traceback.
A simple test case:

request = requests.Request(method='GET', url='https://httpbin.org/get')
request.data =  u'x=føø'
prepared = request.prepare()
requests.Session().send(prepared)

Traceback (most recent call last):
...
    requests.Session().send(prepared)
  File "/usr/lib/python2.7/dist-packages/requests/sessions.py", line 566, in send
    r = adapter.send(request, **kwargs)
  File "/usr/lib/python2.7/dist-packages/requests/adapters.py", line 331, in send
    timeout=timeout
  File "/usr/lib/python2.7/dist-packages/urllib3/connectionpool.py", line 558, in urlopen
    body=body, headers=headers)
  File "/usr/lib/python2.7/dist-packages/urllib3/connectionpool.py", line 383, in _make_request
    conn.request(method, url, **httplib_request_kw)
  File "/usr/lib/python2.7/httplib.py", line 975, in request
    self._send_request(method, url, body, headers)
  File "/usr/lib/python2.7/httplib.py", line 1009, in _send_request
    self.endheaders(body)
  File "/usr/lib/python2.7/httplib.py", line 971, in endheaders
    self._send_output(message_body)
  File "/usr/lib/python2.7/httplib.py", line 835, in _send_output
    self.send(message_body)
  File "/usr/lib/python2.7/httplib.py", line 805, in send
    self.sock.sendall(data)
  File "/usr/lib/python2.7/ssl.py", line 329, in sendall
    v = self.send(data[count:])
  File "/usr/lib/python2.7/ssl.py", line 298, in send
    v = self._sslobj.write(data)
UnicodeEncodeError: 'ascii' codec can't encode characters in position 3-4: ordinal not in range(128)

When data is a list, it works fine.

    request = requests.Request(method='GET', url='https://httpbin.org/get')
    request.data = [(u'x', u'føø')]
    prepared = request.prepare()
    requests.Session().send(prepared)

Requests seems to be supporting unicode data, shown in the code. However, it seems not work well, and the behavior differs when the data is a list-like type and a text-like type, one try to encode using 'uft-8' and the other just pass through(in _encode_params).

    def prepare_body(self, data, files, json=None):
            ...
            if files:
                (body, content_type) = self._encode_files(files, data)
            else:
                if data and json is None:
                    body = self._encode_params(data)     #===> check data
                    if isinstance(data, basestring) or hasattr(data, 'read'):
                        content_type = None
                    else:
                        content_type = 'application/x-www-form-urlencoded'
                    ...
    @staticmethod
    def _encode_params(data):
        if isinstance(data, (str, bytes)):   #===> allows unicode
            return data                              #===> simply return
        elif hasattr(data, 'read'):
            return data
        elif hasattr(data, '__iter__'):        #===> the behavior is different.
            result = []
            for k, vs in to_key_val_list(data):
                if isinstance(vs, basestring) or not hasattr(vs, '__iter__'):
                    vs = [vs]
                for v in vs:
                    if v is not None:
                        result.append(
                            (k.encode('utf-8') if isinstance(k, str) else k,
                             v.encode('utf-8') if isinstance(v, str) else v))   #===>you try to encode as 'utf8'
            return urlencode(result, doseq=True)
        else:
            return data

Start with Request.prepare_body and _encode_params, then reproduce the reported Python 2 behavior using the scalar Unicode and list-of-tuples examples. Compare how each input is encoded and sent, and add regression coverage for the chosen consistent behavior. Done means Unicode request data no longer raises the reported encoding error and the existing list form remains covered.