Problem
Research questions and friction points this paper is trying to address.
file-level issue localization
repository-grounded repair
automated program repair
LLM repair
repair evaluation
Loc2Repair: A Framework for Evaluating the Impact of File-Level Issue Localization in Repo-Level LLM Repair
🤖 AI Summary
This work addresses a critical limitation in existing warehouse-scale large language model–based program repair methods, which often conflate bug localization with patch generation, thereby obscuring the true source of performance bottlenecks. To remedy this, we propose Loc2Repair, a novel framework that systematically decouples file-level localization from the repair phase for the first time, enabling independent quantification of localization’s impact on repair efficacy within a unified evaluation protocol. Through a modular architecture that integrates diverse localization models with various repair backbones, controlled experiments on SWE-bench Verified demonstrate that explicit localization—particularly with oracle (gold) labels—boosts repair success rates from 44.7% to 52.4% and reduces average latency by up to 154.45 seconds. These findings establish localization as an efficient and pivotal lever for effective repair while simultaneously highlighting substantial room for improvement beyond localization alone.
📝 Abstract
Repository-grounded automated repair is often reported as a single end-to-end capability, which hides distinct failure modes such as poor file targeting, incorrect patch synthesis, and failed iterative debugging. We present Loc2Repair, a modular evaluation framework for controlled analysis of repository-grounded repair pipelines, and use it to isolate file-level issue localization as an upstream variable. Loc2Repair decouples localization and repair under a shared runtime, artifact schema, and evaluation harness, allowing researchers to combine different localization models and repair backbones under matched conditions. Using three repair backbones on SWE-bench Verified, we compare baseline repair without explicit localization, repair guided by predicted localization from two localizers, and repair guided by gold modified-file sets. Explicit localization consistently improves resolved rate across all backbones: pooled performance increases from 44.7% for baseline repair to 48.9% and 49.1% with predicted localization, and to 52.4% with gold localization. Localization also reduces mean elapsed time overall: in pooled paired analysis, mean elapsed time decreases by 100.94 s and 52.25 s for the two predicted-localization settings, and by 154.45 s with gold guidance, although token effects remain heterogeneous across models. Overall, Loc2Repair shows file-level localization is a consistent repair lever, improving effectiveness and mean latency in pooled analysis, while gold-guided failures expose headroom beyond localization.
Innovation
Methods, ideas, or system contributions that make the work stand out.
file-level localization
repository-grounded repair
modular evaluation framework
LLM-based program repair
SWE-bench
🔎 Similar Papers
No similar papers found.
