Menu
In testing web applications, details of user visits may be recorded in a web log and converted to test cases. This is called user-session-based testing and studies have shown that such tests may be effective at revealing faults. However, for popular web applications with a larger user base, many user-sessions may build up and test suite management techniques are needed. In this paper, we focus on the problem of test suite prioritization. That is, given a large test suite, reorder the test cases according to a criterion that is hypothesized to increase the rate of fault detection. Previous work shows that 2-way combinatorial-based prioritization is an effective prioritization criterion. We develop a greedy algorithm where we consider memory usage and the time that it takes to prioritize test suites. We represent software tests in a graph by storing unique parameters as vertices and \( n \)-way sets as edges or series of edges. Our experiments demonstrate the efficiency of this approach.