Unveiling the Relationship Between Continuous Integration and Code Review: A Study with 10 Closed-source Projects

Companies have adopted Continuous Integration (CI) and Code Review (CR) as key practices to monitor and improve software quality continuously. These practices enable early detection and correction of issues and have shown promising results in open and closed-source projects. However, there is limited understanding of the relationship between CI and CR in closed-source environments. For instance, it remains unclear how CI practices influence CR (and vice versa) and to what extent bad CI practices hinder the CR process. To address this gap, we conducted a mixed-methods study involving ten closed-source projects developed by industry partners. We collected and analyzed 32 metrics related to CI and CR. In addition to the quantitative analysis, we gathered perceptions from project teams to (i) validate the identified correlations, (ii) explore the effects of bad CI practices on CR, and (iii) understand the perceived benefits and challenges of using CI and CR together. Our results revealed 23 correlations between CI and CR processes, leading to the following key findings: (i) the workload and execution time of CI can influence code review time; (ii) CI and CR execution times are more correlated when the processes occur sequentially; (iii) bad CI practices (such as long-lived branches, poor testing strategies, or complex build schemes) can negatively impact CR, causing delays, reviewer overload, and undetected issues; (iv) CI adds value to CR by anticipating problems, reducing manual tasks, and supporting the distribution of updates; and (v) the joint use of CI and CR presents challenges, such as ensuring code quality, resolving merge conflicts, and aligning processes. These findings shed light on the interplay between CI and CR and offer insights to improve their combined use in closed-source software development.