<scp>iBiR</scp> : Bug-report-driven Fault Injection

Much research on software engineering relies experimental studies based fault injection. Fault injection, however, is not often relevant to emulate real-world faults since it “blindly” injects large numbers of faults. It remains indeed challenging inject few but realistic that target a particular functionality in program. In this work, we introduce iBiR , injection tool addresses challenge by exploring change patterns associated user-reported To faults, create mutants re-targeting bug-report-driven automated program repair system, i.e., reversing its code transformation templates. further appealing practice requires deep knowledge neither nor tests, just the program’s bug reports. Thus, our approach focuses feature targeted report. We assess considering Defects4J dataset. Experimental results show outperforms performed traditional mutation testing terms semantic similarity with original bug, when applied at either system or class levels granularity, and provides better, statistically significant estimations test effectiveness (fault detection). Additionally, injecting 100 couple real ones around 36% cases, while achieves less than 4%.