DepWareTrans: Dependency-Aware Incremental Repository Migration across Co-executable Languages

2026-08-14

AI Summary

This study addresses the failure of repository-level code migration caused by large language models overlooking cross-file dependencies. We propose a dependency-aware incremental migration framework that transcends single-file limitations by constructing dependency graphs to group translation units into dependency-consistent batches. This approach integrates compilation- and test-driven iterative verification to ensure semantic consistency. Evaluated on an industrial system comprising 51,000 lines of code, the framework achieves 100% pass rates for both compilation and testing. It significantly outperforms traditional file-level methods with faster convergence, effectively resolving critical challenges regarding completeness and scalability in large-scale code migration tasks.

Abstract

Repository-level code translation is critical for modernizing legacy systems, yet existing approaches based on large language models (LLMs) operate at the file level and fail to scale to codebases with complex inter-file dependencies. This limitation is evident in our industrial setting, where we aim to migrate a production repository (STAR) from Java to Kotlin, but file-level approaches produce fragmented results and fail to achieve end-to-end correctness. In this paper, we show that the primary cause of failure at the repository level is dependency inconsistency. Through an empirical study on open-source and industrial systems, we find that most errors arise from unresolved cross-file dependencies that cannot be effectively addressed by iterative feedback alone. We propose a dependency-aware incremental migration framework that elevates the unit of translation from individual files to dependency-consistent batches. Our approach constructs a dependency graph, groups interdependent files, and performs batched translation with iterative compile- and test-driven validation. We evaluate our method on a 51K line of code (LOC) industrial system and multiple repositories across interoperable language pairs (Java-Kotlin, Java-Scala, and C#-F#). On the STAR repository, file-level approaches achieve 38.16% compilation and 9.39% test success, whereas our approach achieves 100% compilation and test success across the evaluated settings, converging within a small number of iterations. These results show that dependency-aware batching improves scalability and reliability in repository-level code translation.

Problem

• Repository-level code translation
• Cross-file dependencies
• Dependency inconsistency
• Legacy system modernization

Innovation

• Dependency-Aware Translation
• Incremental Repository Migration
• Batched Translation
• Compile- and Test-Driven Validation
• Cross-file Dependencies