Gerrit relation chains slow merges by 2.6x, study finds
Dependency chains in code review take 2.6x longer to merge than solo changes.
Researchers from Concordia University and Polytechnique Montreal analyzed 29,580 relation chains across 15 repositories in three Gerrit-based ecosystems (OpenStack, Wikimedia, and ONAP), encompassing 401,256 code changes. Their findings show that relation chain usage is growing, with prevalence ranging from 5% to 49% across projects and increasing in 14 out of 15. Chain changes take a median of 2.6 times longer to merge than size-matched solo changes, and the gap widens for very large changes. Review effort propagates through dependency links, with base-change activity correlating with descendant review activity (rho = 0.43–0.61).
Additionally, 33.5% of chain members undergo structural evolution during review, meaning developers frequently restructure dependencies after submission. The authors conclude that relation chains act as durable, ecosystem-shaped coordination units that traditional change-centric analyses overlook. They recommend that future review analytics, reviewer-assignment systems, and AI-assisted review tools should reason over chains rather than isolated changes to better capture review dynamics and reduce overhead.
- Chain prevalence ranges from 5% to 49% across projects and is increasing in 14 of 15 repositories.
- Chain-linked changes take a median of 2.6x longer to merge than size-matched solo changes.
- 33.5% of chain members undergo structural evolution during the review process.
Why It Matters
Review tools must account for dependency chains to improve merge efficiency and reduce coordination overhead.