(B I/AI) Integrating Test-to-Code Trace Link Recovery Approaches

Test-to-code trace links establish a connection between parts of some source code and test cases that assess the behavior of the code. These links are especially beneficial for program comprehension, because tests show by concrete examples how parts of the code are supposed to be used. As manually deriving trace links is time-consuming, research in the field has proposed various automated techniques. The tool TCTracer [1,2] combines multiple recovery techniques for the programming language Java. However, these techniques make strong assumptions about test naming or structure and, thus, rely heavily on developer discipline.

Our research has addressed this limitation through Mut4T, a novel mutation-based [3] approach that establishes trace links without syntactic constraints. The goal of this thesis is to integrate Mut4T into the tool TCTracer [2] so as to make it easily accessible, and  provide a more robust method for connecting tests and production code.

In addition, the thesis shall evaluate the ability of TCTracer combined with Mut4T for recovering test-to-code trace links based on an existing benchmark.

The ideal student for this thesis topic should have good programming skills in Java and a strong interest in software testing.