Guided pattern mining for API misuse detection by change-based code analysis

Abstract Lack of experience, inadequate documentation, and sub-optimal API design frequently cause developers to make mistakes when re-using third-party implementations. Such misuses can result in unintended behavior, performance losses, or software crashes. Therefore, current research aims automatically detect such by comparing the way a developer used an previously inferred patterns correct usage. While has made significant progress, these techniques have not yet been adopted practice. In part, this is due lack process capable seamlessly integrating with development processes. Particularly, existing approaches do consider how collect relevant source code samples from which infer patterns. fact, collection usage pattern miners irrelevant leads false alarms instead finding true misuses. paper, we target problem (a) providing method that increases likelihood true-positive concerning given set changes agnostic concrete static, intra-procedural mining technique (b) introducing concept for just-in-time misuse detection analyzes at time commit. introduce different, lightweight search filtering strategies evaluate them on two real-world datasets determine their usefulness Our main results are (1) commit-based subsequent effectively decreases amount be analyzed, (2) particular method-level superior file-level filtering, (3) project-internal project-external find solutions different types thus complementary, (4) incorporating prior knowledge misused into negligible effect.