How well do test case prioritization techniques support statistical fault localization

Abstract

In continuous integration, a tight integration of test case prioritization techniques and fault-localization techniques may both expose failures faster and locate faults more effectively. Statistical fault-localization techniques use the execution information collected during testing to locate faults. Executing a small fraction of a prioritized test suite reduces the cost of testing, and yet the subsequent fault localization may suffer. This paper presents the first empirical study to examine the impact of test case prioritization on the effectiveness of fault localization. Among many interesting empirical results, we find that coverage-based techniques and random ordering can be more effective than distribution-based techniques in supporting statistical fault localization. Furthermore, the integration of random ordering for test case prioritization and statistical fault localization can be effective in locating faults quickly and economically. © 2009 IEEE.