Assessing, generating, and prioritizing tests for asynchronous JavaScript programs

点击翻译按钮获取中文摘要

The web has had a significant impact on many aspects of our lives and societies. Web applications have become integral parts of our education, entertainment, governments, businesses, and personal and social lives. Rigorous testing is required to ensure all these applications function as expected. However, testing web applications is a particularly challenging endeavour. JavaScript has been the lingua franca of client-side web development. More recently, platforms such as Node.js have made it possible to use JavaScript for writing code that runs outside of the browser on a light-weight, fast, and scalable platform. As such, server-side JavaScript development has also experienced an exponential growth in the past few years. JavaScript owes some of its popularity to the particular set of the language's paradigms, such as asynchronous execution. JavaScript is single-threaded and thus asynchronous callbacks are used regularly to simulate concurrency and provide capabilities such as rich user interactions, non-blocking I/O, and seamless client/server communications. Unfortunately, the analysis and testing techniques have not evolved nearly as much as our programming paradigms in web development have, and are thus not sufficient for detecting and revealing various bugs in asynchronous code. The goal of this research is to facilitate the process of testing asynchronous JavaScript programs in order to allow the development of robust web applications. To this end, we will provide novel techniques and tool support in development environments for JavaScript programmers and testers' everyday testing activities. Our proposed research will examine traditional testing techniques to better understand and address their shortcomings in the presence of new programming paradigms. We propose to revisit testing of asynchronous code, by introducing new paradigms for assessing the adequacy of existing testing techniques for asynchronous JavaScrip execution, generation of tests that address inadequacies of traditional testing for this evolved programming paradigm, and prioritization of the tests. The outcome of this research will have great impact in research and in industry, and will evolve our view of testing asynchronous code.