July 2010 Summaries

Monocle is an async programming framework developed by Greg and Steven Hazel from Sauce Labs, which aims to provide a blocking look-alike syntax for event-driven I/O frameworks. The framework is designed to be portable between different event-driven I/O frameworks, currently supporting Twisted and Tornado. This development could indicate that Sauce Labs is shifting its focus towards concurrent web performance and addressing the limitations of traditional async programming methods.

Adobe's Flash and Flex are still relevant in the web development landscape despite Apple's support for HTML5. The massive installed base of Flash on 98% of computers ensures its continued use, along with comprehensive tool sets that developers need. While HTML5 has its advantages, it cannot yet match the capabilities of Flash in areas like webcam capture, media-streaming, and full-screen video. The Sauce Labs announcement of the Sauce Flash-Flex Testing System addresses this gap by providing testing of Flex applications on the Flash platform, catering to the large market of Flash-based gaming and advertising.

There has always been some overlap between Hudson users and Selenium users, which has resulted in good support for Selenium in Hudson. However, setting up the right OS/browser combinations can be tedious, especially without virtualization. Sauce OnDemand solves this problem by offering browsers running in the cloud, but setting up SSH port forwarding to handle this requires time-consuming setup and teardown. To fix this, I wrote a new "Sauce OnDemand" plugin for Hudson that automatically sets up and tears down tunnels for users, saving time and complexity. The plugin also auto-generates unique host names when setting up tunnels, allowing concurrent testing without collisions. The implementation is done in two parts: an SSH tunnel setup library and a Hudson plug-in around it, which can be reused elsewhere. This plugin simplifies the process of using Sauce OnDemand with Hudson, making cross-browser testing easier and more efficient.