Efficiently Finding Useless Mutants

Mutation analysis is a popular but costly approach to assess the quality of test suites. Equivalent and redundant mutants contribute to increase costs and are harmful to the design of test suites. We propose a lightweight technique to identify equivalent and redundant mutants based on theorem proving with Z3 in the context of weak mutation testing. The experiments reveal that our technique detects all equivalent mutants detected by TCE and we have an average reduction of 72.52% of mutants, when considering entire programs. We also apply our technique on HOMs. When considering both FOMs and HOMs, 91% of the mutations could be discarded on average. The results found by our approach may help to make mutation testing less expensive and more accurate.