A Searchbased Multi-Objective Approach To Generate Test Suites For High Branch Coverage

Abstract

A software test consists of an input that implements the program and a definition of the expected outcome. Many techniques to automatically create inputs have been proposed over the years and today are competent to produce test suites with high code coverage. Yet the problem of the expected outcome continues and has become known as the oracle problem. To make this feasible test generation needs to intend not only at high code coverage but also at small test suites that make oracle generation as easy as possible. Coverage goals are not sovereign, not evenly difficult and sometimes infeasible. The result of test generation is therefore dependent on the order of coverage goals and how many of them are possible. To overcome this problem we propose a novel paradigm in which whole test suites are developed with the aspire of covering all coverage goals at the same time while keeping the total size as small as possible. This approach has several advantages as for example, its efficiency is not affected by the number of infeasible targets in the code. We have implemented this novel approach in the EVOSUITE tool and evaluated it to the frequent approach of addressing one goal at a time.