Refactoring Practices in the Context of Modern Code Review: An Industrial Case Study at Xerox

TL;DR

This industrial case study investigates how professional developers at Xerox review, perceive, and document refactoring activities during modern code review, highlighting challenges and proposing improved documentation procedures.

Contribution

The paper provides novel insights into refactoring review practices in industry and proposes a structured documentation procedure to address identified challenges.

Findings

Lack of proper documentation procedures for refactoring during review

Reviewers face difficulties understanding refactoring intent and implications

Proposed documentation procedure aims to improve review clarity and effectiveness

Abstract

Modern code review is a common and essential practice employed in both industrial and open-source projects to improve software quality, share knowledge, and ensure conformance with coding standards. During code review, developers may inspect and discuss various changes including refactoring activities before merging code changes in the codebase. To date, code review has been extensively studied to explore its general challenges, best practices and outcomes, and socio-technical aspects. However, little is known about how refactoring activities are being reviewed, perceived, and practiced. This study aims to reveal insights into how reviewers develop a decision about accepting or rejecting a submitted refactoring request, and what makes such review challenging. We present an industrial case study with 24 professional developers at Xerox. Particularly, we study the motivations, documentation practices, challenges, verification, and implications of refactoring activities during code review. Our study delivers several important findings. Our results report the lack of a proper procedure to follow by developers when documenting their refactorings for review. Our survey with reviewers has also revealed several difficulties related to understanding the refactoring intent and implications on the functional and non-functional aspects of the software. In light of our findings, we recommended a procedure to properly document refactoring activities, as part of our survey feedback.