Explicitly set encoding for http requests in tests

Due to the upgrade to charset-normalizer 2.0.4 guessing the encoding
inside the tests did not work anymore and caused the umlaut tests to
fail. Explicitly specifying the encoding on the requests' response
object fixes this.

tests/test_servefile.py

@@ -92,11 +92,14 @@ def datadir(tmp_path):
 
 
 def make_request(path='/', host='localhost', port=SERVEFILE_DEFAULT_PORT, method='get', protocol='http', timeout=5,
-                 **kwargs):
+                 encoding='utf-8', **kwargs):
     url = '{}://{}:{}{}'.format(protocol, host, port, path)
     print('Calling {} on {} with {}'.format(method, url, kwargs))
     r = getattr(requests, method)(url, **kwargs)
 
+    if r.encoding is None and encoding:
+        r.encoding = encoding
+
     return r