JavaScript Instrumentation for Search-Based Software Testing: A Study with RESTful APIs

JavaScript is one of the most popular programming languages. However, its dynamic nature poses several challenges to automated testing techniques. In this paper, we propose an approach and open-source tool support to enable white-box testing of JavaScript applications using Search-Based Software Testing techniques. We provide an automated approach to collect search-based heuristics like the common Branch Distance. To empirically evaluate our results, we integrated our technique into the EvoMASTER test generation tool, and carried out analyses on the automated system testing of RESTful APIs. Experiments on 5 NodeJS APIs show that our technique leads to significantly better results than existing black-box and grey-box testing tools in terms of code coverage and fault detection.