Prediction of Bug Inducing Commits Using Metrics Trend Analysis

Continuous software engineering advocates a release-small, release-often process model, where new functionality is added to a system very frequently and in small increments. In such a process model, it is important to be able to identify as early as possible, and every time a change is introduced, whether the system has entered a state where faults are more likely to occur. In this paper, we present a method that is based on process, quality, and source code metrics to evaluate the likelihood that an imminent bug inducing commit is highly probable. More specifically, the method analyzes the correlations, and the rate of change of selected structural and quality metrics. The findings from the SonarQube Technical Debt open-source dataset indicate that before bug inducing commits, metrics which otherwise are not corelated, suddenly exhibit a high correlation or high rate of metric value change. This metric behavior can then be used as a predictor for a imminent bug inducing commit. The technique is programing language agnostic, as it is based on metrics which are extracted without the use of specialized parsers, and can be applied to forewarn developers that a file, or a collection of files, has entered a state where faults are highly probable.