Go to ICLR 2025 Conference homepageGenerate-then-Test: Automated Test Case Generation for WebAssembly Using Large Language Models
Keywords: WebAssembly, Large Language Models, Test Case Generation
TL;DR: We present WasmTest, an automated tool that uses LLMs to generate and mutate test cases for WebAssembly binaries, enhancing bug detection and code coverage without needing source code.
Abstract: The reliability and security of WebAssembly (Wasm) binaries are crucial for modern web development, yet effective testing methodologies remain undeveloped. This paper addresses the gap in Wasm binary testing by proposing a novel approach for test cases generation, leveraging Large Language Models (LLMs) to enhance test coverage and bug detection. Traditional testing approaches typically require access to source code, which is often unavailable for Wasm binaries. Our generate-then-test methodology overcomes this limitation by generating equivalent C++ code to simulate expected Wasm behavior, creating and mutating test cases in C++, and compiling these tests to evaluate them against the Wasm binary. Key contributions include automated test case generation using LLMs and improved code coverage through type-aware mutations, with comprehensive evaluation demonstrating the effectiveness of our approach in detecting subtle bugs in Wasm binaries, thereby ensuring more reliable Wasm applications.
Primary Area: applications to computer vision, audio, language, and other modalities
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 9482
Loading