ACE: Self-Evolving LLM Coding Framework via Adversarial Unit Test Generation and Preference Optimization

Collection of recent CS/SE papers focused on LLM-driven code generation, testing, tooling, and governance. Highlights include ACE, a self‑evolving solver–adversary LLM framework that generates adversarial unit tests to drive improvement; enterprise-scale LLM adaptation (Gemini for Google) and operationalization guidance; TARIPlay for automated AR app testing from playback videos; an empirical comparison of LLM-based vs. search-based merge conflict resolution; IDE and tooling work for the Move smart-contract ecosystem; an empirical study of GitHub repository contents and trends (CI, config, LLM