DepRepair: LLM-Based Source-Code Repair for Dependency Breaking Changes

2026-07-20

AI Summary

This work addresses the challenge of breaking changes introduced by third-party library updates, which often cause downstream code failures and are inadequately handled by existing repair methods due to their limited use of cross-repository upstream evidence. To tackle this, we propose DepRepair, a large language model–based one-shot repair approach that leverages structured upstream evidence—such as release notes and API diffs—to guide the repair process. We introduce DepBench, the first cross-repository benchmark for dependency-breaking repairs, and design three key components: evidence filtering, usage localization, and subclass-aware prompting, enabling precise, breakage-type–sensitive fixes. Experimental results show that DepRepair achieves executable pass rates of 89.5% and 82.1% on DepBench when integrated with GPT-5.5 and Claude Opus 4.6, respectively, significantly outperforming baselines and demonstrating the critical role of structured evidence in enhancing repair effectiveness.

Abstract

Modern software projects depend on numerous third-party libraries, whose updates often introduce breaking changes. Adapting consumer code to such changes remains labor-intensive and error-prone. Existing work either characterizes dependency breaking changes without producing a verified consumer-side patch, or studies automated repair only in settings where the failure and repair context are contained within the target repository. However, dependency breaking changes violate this assumption: the decisive repair evidence lies upstream in release notes and API diffs, and no failing test localizes where the consumer breaks, leaving the repair under-informed. To study this cross-repository problem on real data, we introduce DepBench, a benchmark of 95 real-world dependency-update instances across four ecosystems, each paired with a Docker-based executable oracle that runs the consumer's own tests. To address these challenges, we propose DepRepair, a single-call LLM approach that grounds repair in structured upstream evidence through three components: an evidence filter that distills relevant upstream changes, a usage locator that identifies affected consumer sites, and a subcategory-aware guide that tailors repairs to the breaking-change type. Evaluated on DepBench, DepRepair attains the highest executable pass rate on each backbone, achieving 89.5% with GPT-5.5 and 82.1% with Claude Opus 4.6. We further find that raw upstream evidence reduces LLM and agent pass rates by 7--23 percentage points, whereas structured evidence consistently improves them.

Problem

• dependency breaking changes
• source-code repair
• cross-repository repair
• automated program repair
• LLM-based repair

Innovation

• LLM-based program repair
• dependency breaking changes
• structured evidence grounding
• cross-repository repair
• DepBench