Remove @pytest.mark.fuzz

This was an unregistered pytest mark, and pytest now emits a warning,
which causes our tests to fail, when an unregistered mark is used:

pytest-dev/pytest#4935

Rather than registering the mark simply delete all uses of it. All tests
that're defined using hypothesis automatically have the
`@pytest.mark.hypothesis` mark applied to them so it's not necessary to
try to remember to manually apply a custom mark to every test that uses
hypothesis:

https://hypothesis.readthedocs.io/en/latest/details.html#the-hypothesis-pytest-plugin

Author: seanh

tests/h/auth/tokens_test.py

@@ -77,7 +77,6 @@ def test_returns_none_for_malformed_header(self, pyramid_request):
         assert result is None
 
     @given(header=st.text())
-    @pytest.mark.fuzz
     def test_returns_none_for_malformed_header_fuzz(self, header, pyramid_request):
         assume(not header.startswith("Bearer "))
         pyramid_request.headers["Authorization"] = header

tests/h/search/parser_test.py

@@ -128,7 +128,6 @@ def test_parse_with_odd_quotes_combinations(query_in, query_out):
 
 @given(st.text())
 @settings(max_examples=30)
-@pytest.mark.fuzz
 def test_parse_always_return_a_multidict(text):
     """Given any string input, output should always be a MultiDict."""
     result = parser.parse(text)
@@ -144,7 +143,6 @@ def test_parse_always_return_a_multidict(text):
 
 @given(kw=st.sampled_from(parser.named_fields), value=nonwhitespace_text)
 @settings(max_examples=30)
-@pytest.mark.fuzz
 def test_parse_with_any_nonwhitespace_text(kw, value):
     result = parser.parse(kw + ":" + value)
     assert result.get(kw) == value