Slice-based statistical fault localization

Deep Learning-Based Fault Localization with Contextual Information

“…The 6 small-scale subject programs and 4 large-scale subject programs are all widely used in the fault localization community. Our experiments adopt widely used metrics EXAM [8] and RImp [12], the former is defined as the percentage of executable statements to be examined before finding the actual faulty statement, the latter is to compare the total number of statements that need to be examined to find all faults by our approach versus the number that need to be examined by using other fault localization technique. Lower values of both metrics show better improvement.…”

“…Another drawback of many techniques is that they ignore the contextual information of the relationship among suspicious statements [4]. Hence, program slicing technology could be taken into consideration for its ability of extracting data and/or control dependencies of program statements and selecting a subset of statements affecting the incorrect output [7], [8]. There are static slicing and dynamic slicing, where dynamic slicing gathers run-time information along the execution path and cut down the size of the slice noticeably in comparison with static slicing [9].…”

The Impact of Information Richness on Fault Localization

“…The details of SFL using BC, SFL using FC and SFL using BS can refer to [1], [4] and [9], respectively. First, we assume that a program P comprises a set of program statements S = {s 1 , s 2 , .…”

“…1). Let bslice(t i ) be the backward slice [9] of the output of the test case t i . Thus, bslice(t i ) includes those statements whose execution affects the output of test case t i according to data and control dependence.…”

“…The experiment first chooses three major information types used by SFL. The three types of information are BC [1], FC [4] and Backward Slice * (BS [9]). In addition, we follow the structure of SFL to construct a new SFL technique using a new information type that combines FC with BS (FC&BS).…”