The MarkDecorator __eq__ method assumes that it will always be compared with another mark. If you do large parametrizations based on lists, then you'll likely run into scenarios where this isn't the case (especially if you do combinations of lists).

Simple example:

import pytest
TEST_DATA = ['a', 'b', 'c', pytest.mark.xfail('d')]
SUBSET_TESTS = [x for x in TEST_DATA if x in ('a', 'b')]

@pytest.mark.parametrize('data', SUBSET_TESTS)
def test_me(data):
     pass

You'll get an error on collection:

    return self.mark == other.mark
E   AttributeError: 'str' object has no attribute 'mark'

This affects sets, lists (including list.index!), anything that checks for equality.