Name: Selenium with Python and pytest
Availability: InStock

Ten questions on the decisions this course made and why. Each answer explains the reasoning once you check. The quiz is a self-check; nothing is recorded.

1. Code copied from an old tutorial calls driver.find_element_by_id("email") and fails on Selenium 4.52 with AttributeError. What is the fix?

Downgrade Selenium until the method exists again. Use driver.find_element(By.ID, "email"); the find_element_by_* helpers were removed in Selenium 4. Install a newer chromedriver.

Selenium 4 replaced the per-strategy helpers with find_element(By.X, value), and 4.3 removed the old ones. Downgrading locks you out of current browsers, and the driver binary has nothing to do with the Python API.

2. You add tests/login_checks.py with a function test_login, but pytest reports "no tests ran". Why?

By default pytest only collects files named test_*.py or *_test.py. The function needs a @pytest.mark.test decorator. Test files must be imported from conftest.py.

Discovery is name-based: rename the file to test_login.py. pytest needs no decorator to find a test_ function, and nothing should import test files.

3. wait.until(EC.element_to_be_clickable(By.ID, "save")) raises TypeError. What is wrong?

element_to_be_clickable only accepts an element, not a locator. The wait has no timeout. Expected conditions take one locator tuple: EC.element_to_be_clickable((By.ID, "save")).

The condition's single argument is the tuple (By.ID, "save"). Passing two separate arguments is a Python TypeError before any waiting happens. The condition accepts either a locator tuple or an element.

4. A fixture sets driver.implicitly_wait(10) and the tests use WebDriverWait(driver, 15). Some failures take around 20 seconds to report. Why?

The grid adds a fixed 5-second delay to every command. Mixing implicit and explicit waits makes timing unpredictable, because element lookups inside the explicit wait also wait implicitly. WebDriverWait always rounds up to the next 10 seconds.

Each poll of the explicit wait can block for the implicit timeout, so the combined wait overshoots. The Selenium docs give this exact example. Keep the implicit wait at 0 and use explicit waits.

5. In a yield fixture, the test body raises an AssertionError. What happens to the code after yield?

It runs, so the browser is still closed. It is skipped, so you need try/finally in every test. It runs only when pytest is started with --teardown.

pytest runs fixture teardown after the test regardless of the outcome. That is why the fixture can replace the try/finally from the first chapter. Teardown is skipped only if the fixture's own setup failed before reaching yield.

6. A test is decorated with @pytest.mark.parametrize("value", ["a", "b", "c"]) and uses the driver fixture. You run pytest --browser chrome --browser firefox. How many test items does it produce?

3 5 6

Parametrisations multiply: three values times two browsers. Each combination is a separate test with its own ID, such as [firefox-b], and its own browser.

7. Why does the course's conftest.py implement pytest_runtest_makereport?

To generate the HTML report. To store each phase's report on the test item, so the fixture's teardown can tell whether the test failed. To retry failed tests automatically.

Fixtures do not receive the test outcome. The hook saves it as rep_call on the item, and the fixture reads request.node.rep_call to decide on a screenshot and the grid status. HTML reports come from pytest-html, and the hook retries nothing.

8. In the page object, locators are tuples such as SUBMIT = (By.CSS_SELECTOR, "button[type='submit']"). Why write find_element(*self.SUBMIT)?

The star unpacks the tuple into the two arguments find_element expects, and the same tuple can be passed whole to expected conditions. The star makes Selenium wait for the element. The star finds all matching elements instead of the first.

* is Python argument unpacking: find_element(*("css selector", "...")) is find_element("css selector", "..."). It adds no waiting, and find_elements is how you get all matches.

9. A test file keeps a module-level list of the users its tests created, and a final test asserts the list has five entries. It passes with -n 0 and fails with -n 4. Why?

pytest-xdist clears module variables after every test. Lists are not thread-safe in Python. Workers are separate processes with their own copy of the module, and the tests are spread across them, so no single list sees all five users.

xdist uses processes, not threads, and does not guarantee which worker runs which test or in what order. The real problem is that the last test depends on the others; each test should set up and check its own data.

10. A pull request puts the TestingBot key and secret directly into conftest.py "so CI works". What should you ask for instead?

Move them into pytest.ini, which is not Python code. Read them from TB_KEY and TB_SECRET environment variables, set from CI secrets, and rotate the leaked pair. Base64-encode them before committing.

Anything committed, in any file or encoding, is readable by everyone with repository access and stays in the history. Environment variables filled from CI secrets keep credentials out of the code, and a leaked pair should be replaced.

Where to go next#

Run the suite on more browser and OS combinations with the capabilities generator.
Grow the page objects into a framework with the Page Object Model course.
See the same suite in Java in Selenium with Java and TestNG, or compare a different tool in the Playwright course.
Prepare for interviews with the Selenium interview questions.

Selenium Python Quiz

Where to go next#

You finished Selenium with Python and pytest