py3 compatibility: convert bytes to str (#33206)

Fixes #31723
In python3, ansible's fetch_url utility function wraps urllib.request.urlopen.
For HTTP and HTTPS URLs, this function returns a http.client.HTTPResponse object slightly modified.
Calling .read() on an HTTPResponse object returns bytes (note the docstring fix).

Here, to_native is used to convert the bytestrings returned by fetch_url into unicode strings.
This is necessary because:
  1. Pre python3.6, json.loads requires passing a string, not a bytestring, as its argument
  2. In python3 generally, testing if a string is a substring of a bytestring
     using the 'in' operator will raise a TypeError

see:
- https://docs.python.org/3/library/urllib.request.html#urllib.request.urlopen
- https://docs.python.org/3/library/http.client.html#http.client.HTTPResponse.read
- https://docs.python.org/3/library/json.html#json.loads

lib/ansible/module_utils/urls.py

@@ -941,7 +941,7 @@ def fetch_url(module, url, data=None, headers=None, method=None,
     :kwarg last_mod_time: Default: None
     :kwarg int timeout:   Default: 10
 
-    :returns: A tuple of (**response**, **info**). Use ``response.body()`` to read the data.
+    :returns: A tuple of (**response**, **info**). Use ``response.read()`` to read the data.
         The **info** contains the 'status' and other meta data. When a HttpError (status > 400)
         occurred then ``info['body']`` contains the error response data::