Automatic patch linkage detection in code review using textual content and file location features

Contemporary code review tools are a popular choice for software quality assurance. Using these tools, reviewers able to post linkage between two patches during discussion. Large development teams that use review-then-commit model risk being unaware of linkages. Our objective is first explore how patch impacts the process. We then propose and evaluate models detect based on realistic time intervals. First, we carry out an exploratory study three open source projects conduct impact analysis using 942 manually classified Second, techniques textual file location similarity build detection their performance. The provides evidence latency in notification. show with Alternative Solution (i.e., implement similar functionality) undergoes quicker avoids additional revisions after team has been notified, compared other types. experiments promising recall rates (from 32% 95%), but precision room improvement. Patch promising, likely improvements if practice posting linkages becomes more prevalent. From our implications, this paper lays groundwork future research increase awareness facilitate efficient reviews.