25.13. Code Coverage

Code coverage measures the percentage of code executed by your tests.

Code coverage is a metric used in software testing to gauge how much of your code is being tested by your unit tests. When you run your tests, code coverage indicates what percentage of your source code has been executed. This helps developers identify untested parts of the code. A higher percentage means more of the code is executing as part of the tests, which generally indicates a more thoroughly tested application.

Think of code coverage like a safety inspection for a building. If inspectors check every room, hallway, and window, the building gets a high safety score because all parts have been evaluated for safety. However, if some rooms are overlooked, then the score isn't as high, indicating potential hazards that haven't been checked.

Tools • JaCoCo (Java Code Coverage) • Cobertura • SonarQube

Several tools help measure code coverage in Java applications. JaCoCo is a widely used tool that can provide detailed reports on which parts of your code were executed during testing. Cobertura is another tool that provides coverage reports as well. SonarQube can analyze the overall quality of code, including code coverage metrics, helping teams track their progress over time. Using these tools helps ensure your tests are robust and effective.

Imagine using a fitness tracker to monitor your daily exercise. Just as the tracker provides data on the percentage of exercises completed, such as walking or running, coverage tools give similar insights about which areas of your code have been tested. You can see what’s covered and what needs more attention, just as you’d check which fitness goals are met.

Goal: Aim for high coverage but not 100% blindly. Some code like error logging may not need to be tested.

While aiming for high code coverage is crucial, it is important to note that achieving 100% coverage is not always necessary or beneficial. Certain parts of code, such as error logging, may not require extensive testing since they do not contain logic that directly contributes to the application's functionality. Developers should focus on testing code that contains significant business logic and functionality while being mindful of diminishing returns that come from striving for perfect coverage.

Consider a student preparing for an exam. While it’s important to cover all topics, spending hours memorizing minute details that won't be on the test can be a waste of time. Similarly, achieving high code coverage is valuable, but focusing too much on every single line can divert attention from testing critical functionality effectively.

--