parametrize: truncate long parameters in output in `-v` mode, still display full value in `-vv`

[nixjdm]

AFAICT this ability doesn't exist.

I have a parameterized test that sometimes receives params that are long strings. I would generally prefer verbose output, but the long strings really clutter up the output pretty badly.

For example:

strings_list = [
    # Imagine several of these
    (
        """Imagine a whole paragraph here.""",
        """Imagine a whole paragraph here.""",
    ),
]

pytest.mark.parametrize(["content", "expected"], string_list)
def test_sample(content, expected):
    ....

Testing the above with pytest -v Will print the entirety of the long strings used in string_list. Normally this is just what I want, and the output for each line fits nicely on a single line, or maybe a little spill over to two. With long params though, this creates a mess that makes it more difficult to understand the total results of the test at a glance, and needs a lot more scrolling to get through.

Nonetheless, a hint to the params would be great. I suggest, to not change current behavior, the ability to decorate or otherwise mark specific tests to affect this behavior for just that test.

Alternatives:

• A global setting to change this behavior (a good supplement)
• Actually changing the default behavior of -v. I'd guess this isn't a welcome option.

[Yay295]

I found this code in pytest's tests that I think could be modified to do what you want.

pytest/testing/python/approx.py

Lines 48 to 65 in 326ae0c

	@contextmanager
	def temporary_verbosity(config, verbosity=0):
	original_verbosity = config.getoption("verbose")
	config.option.verbose = verbosity
	try:
	yield
	finally:
	config.option.verbose = original_verbosity
	@pytest.fixture
	def assert_approx_raises_regex(pytestconfig):
	def do_assert(lhs, rhs, expected_message, verbosity_level=0):
	import re
	with temporary_verbosity(pytestconfig, verbosity_level):
	with pytest.raises(AssertionError) as e:
	assert lhs == approx(rhs)

Something like this:

strings_list = [
    # Imagine several of these
    (
        """Imagine a whole paragraph here.""",
        """Imagine a whole paragraph here.""",
    ),
]

@pytest.mark.parametrize(["content", "expected"], strings_list)
def test_sample(content, expected, pytestconfig):
    original_verbosity = pytestconfig.getoption("verbose")
    pytestconfig.option.verbose = 0
    try:
        ....
    finally:
        pytestconfig.option.verbose = original_verbosity