Improving Failure-inducing Changes Identification using Coverage Analysis
Kai Yu, Mengxiang Lin, Jin Chen, Xiangyu Zhang · 2013
Abstract—Delta debugging has been proposed for failureinducing changes identification. Despite promising results, there are two practical factors that thwart the application of DD: large number of tests and misleading false positives. To address the issues, we present a combination of coverage analysis and delta debugging that automatically isolates failureinducing changes. Evaluations on twelve real regressions in GNU software demonstrate both the speed gain and effectiveness improvements. Keywords-software regression; delta debugging; coverage analysis; execution trace; automated debugging; I.