$1.$ Set up a continuous integration server. Include version control for both source code and tests, and populate both with a simple example. Experiment with $$breaking the build$,$ by either introducing a fault into the source code or adding a failing test case. Restore the build.

$2.$ Find a refactoring in some large, existing system. Build tests that capture the behavior relevant to that part of the system. Refactor, and then check that the tests still pass.

$3.$ Repair a fault in an existing system. That is$,$ find the code that needs to change and capture the current behavior with tests. At least one of these tests must fail, thus demonstrating that you found the fault. Repair the fault and check that all of your tests now pass