Evaluating the Strategies of Statement Selection in Automated Program Repair

Automated program repair has drawn significant attention in recent years for its potential to alleviate the heavy burden of debugging activities. Fault localization, which generally provides a rank list of suspicious statements, is necessary for automated repair tools to identify the fault. With such rank list, existing repair tools have two statement selecting strategies for statement modification: suspiciousness-first algorithm (SFA) based on the suspiciousness of statements and rank-first algorithm (RFA) relying on the rank of statements. However, despite the extensive application of SFA and RFA in repair tools and different selecting methods between both strategies, little is known about the performance of the two strategies in automated program repair.