Spock: Up and Running

Chapter 5. Interaction Testing

Testing how the code-under-test deals with collaborators by using mock and stub test doubles is called interaction testing. This is one of the most misunderstood and abused techniques in automated testing. Inappropriate use of mocks and stubs can tightly couple your test code to the implementation details, causing brittle tests and making refactoring difficult because changes in how a unit works can break tests, even though the behavior is still correct. At the same time, mocks and stubs can enable a modular design that uses Inversion of Control (IoC) to define the relationships between units resulting in loosely coupled code that cooperates through well-defined interfaces.

Mocks and Stubs

The terms “mock” and “stub” are used inconsistently in discussions of testing. They are both forms of test double—a stand-in for a real code unit that is a collaborator of the unit under test. “Collaborator” simply means some other object whose methods are invoked by the unit under test. In Spock, and in this book, the terms are used as follows:

Mock

A test double that tracks when its methods are called and the parameters that are passed to them. Tests can then verify that calls were made as expected.

Stub

A test double whose methods will return a predetermined response or take a predetermined action such as throwing an exception. Stubs do not track the calls made to them but can use sequences of predetermined responses or actions to behave differently when ...