Node.js-based Selenium testing

6:30: Dinner & networking

7:00: Announcements

7:10: Speaker

8:45: Lights out!

ABSTRACT

Creating Selenium tests that are stable and scalable can be challenging. When we created our Node.js-based Selenium testing framework, there were many issues we needed to consider, including:

• Coverage-report collection of code that ran in browsers

• Asynchronicity of JavaScript and test-code readability

• Implementation of page-objects and their re-usability for Single-Page Applications

• Utilization of Selenium to stabilize visual regression tests

• Parallelization and coordination of tests with multiple browsers and across frameworks (Mocha, Cucumber, …)

• Aggregation of test results and consistent reporting

Recently, we open-sourced tools that were created to overcome these challenges. I will talk about the issues we faced, what decisions we made, and how we use these tools to test our UI. Apart from infrastructure, this talk will also spend some time on describing steps we took to keep Selenium tests maintainable and scalable.

SPEAKER

Marcel Erz has been a software engineer for the past 16 years, and has created software on a variety of platforms, including desktop, mobile, PLC’s, and the web. He has a passion for testable code-design and introduced QA processes and best practices in multiple companies for which he has worked. Marcel has been using Selenium since 2009, and has written Selenium code in Java, PHP, and Node.js. He has contributed to the PHPUnit Selenium project, and is a major contributor to the Cabbie WebDriver-client. Currently, Marcel is working in the Programmatic Innovations team at Yahoo!