Exit course
Chapter 10 of 10 Beginner 7 min read Updated

Knowledge check

Ten questions on the decisions that matter in a real Robot Framework suite. Each answer comes with a short explanation, so a wrong answer points you back to the right chapter.

1. A team of testers who do not write Python and two developers will maintain a suite covering a web app and its REST API. Why is Robot Framework a reasonable fit here?

Readability for non-programmers plus one syntax and report across layers is the framework's strength. It is not the only way to test web and API together, it is not faster, and browsers always need a library such as SeleniumLibrary or Browser.

2. A CI agent runs the suite with a globally installed robotframework-browser that someone upgraded last week. Since then every Browser test fails on its first keyword. What setup would have avoided this?

Pinned versions in a per-project environment make upgrades deliberate, and the Browser library's Node side must be cleaned and initialised again after every upgrade. The library does not use a separately installed Playwright.

3. After clicking "Save", a SeleniumLibrary test checks Page Should Contain Saved. It passes on your laptop and fails about one run in five in CI. What is the right fix?

Should keywords check once, so the library timeout does not affect them. The Wait Until variant retries until the text appears and finishes as soon as it does. A sleep only guesses, and reruns hide the problem.

4. Ten test files each contain the same three lines that log in with locators. The login form's markup changes. What structure would have made this a one-line fix?

A user keyword in a resource file keeps the how in one place; tests only say what they do. One big file still repeats the steps, and the locator type does not remove the duplication.

5. The suite should run on Chrome locally and on Safari on TestingBot in CI, without editing any file. Which mechanism does that?

Command-line variables override those defined in files, so the same suite runs anywhere and the choice lives in the command. Copies drift apart, and hostname checks spread environment logic through the tests.

6. All tests pass in Robot's report, but in the TestingBot dashboard none of the sessions shows passed or failed. What is missing?

The grid only sees browser commands, never your assertions, so the result has to be reported. The build name only groups sessions; xunit output is for CI; the tunnel is for private sites.

7. A Browser library suite builds the TestingBot capabilities with &{tb}= Create Dictionary key=%{TB_KEY} secret=%{TB_SECRET}. The tests pass. What is the problem?

Robot Framework logs variable assignments at the default log level. Building the capabilities and connecting inside a Python keyword keeps the secret out of every result file.

8. Twelve rows of a templated login test cover invalid inputs. One row's expected value should be an empty string, but the test fails with "expected 3 arguments, got 2". What fixes it?

Cells are separated by runs of spaces, so an empty cell disappears no matter how many spaces you leave. ${EMPTY} passes an explicit empty string.

9. Your TestingBot plan allows 4 parallel sessions. A colleague sets pabot --processes 16 in CI to finish faster. What happens?

The plan's parallel limit decides how many browsers run. Extra processes only queue, and a queued Browser library connection still counts against its connect timeout. Match --processes to the plan.

10. The GitHub Actions job fails on a Robot test, but the uploaded results artifact is missing. Which step is wrong?

GitHub Actions skips later steps after a failure unless they opt in. Robot always writes the log and report; the upload must run regardless of the result.

Where to go next#

  • Take the TestingBot Robot Framework certification: an exam on the topics of this course, plus a practical task in which your suite runs on the TestingBot grid in a build named certificate-robot-framework.
  • Testing native apps? The Appium and Robot Framework guide uses the same structure with AppiumLibrary, and the mobile app testing course covers the strategy.
  • The page object model and flaky tests courses go deeper into the two topics that decide whether a suite stays maintainable.

You finished Robot Framework Tutorial for Web Testing

Put what you learned into practice. Run your tests on 6100+ real browsers and physical iOS and Android devices.