Automatic Unit-Test Generation Based on Symbolic Execution
Testing is one of the crucial stages of development. Code without proper test coverage leads to undetected defects and vulnerabilities that result in errors or exploitation of critical system components. Nevertheless, providing proper test coverage is a resource-intensive process. This paper introduces a method for automated unit test generation, alongside its functional implementation. The tool accepts the source code and the target function name as input. It then symbolizes the function's parameters and the return values of function calls. Afterwards, it performs symbolic execution of the target function, processes the result, recovering symbolic pointer connections to the underlying objects, restoring their offsets. It creates mock functions, which simulate the behaviour of the real ones, and creates a main function for each execution path. The generated file will be the unit test exploring that path. Finally, it outputs a file that contains the target function and a unit test file for each execution path. The implemented tool runs automatically, without the need for manual configuration.