November 2012 Summaries

Running Automated Tests in Parallel (Part 1)` provides insights into maximizing the value of automated functional tests by providing results quickly to developers. Typically, test suites can take many hours to run due to sequential execution and data dependence. To overcome this, batches of tests can be created and run on separate machines with their own instances of the application under test and database, utilizing spare hardware or cloud computing resources. Continuous Integration servers like Jenkins and Bamboo enable easy setup for running builds on multiple computers, known as Slaves or Agents, allowing grouping of tests into categories or groups that can then be executed in parallel. This enables faster feedback to developers, reducing the likelihood of issues remaining unaddressed.

The Eschaton` refers to the end of time in theological circles, but in the context of testing with Selenium, it represents a mature testing environment. Visualizing the end result of what you're building before laying down any foundation can lead to better decision-making. A "zombie readiness kit" has been devised to help set up this environment, consisting of five key components: storing tests in a source code repository, running tests on a service like Sauce Labs or a Selenium server, triggering tests with a continuous integration server like Jenkins, logging and tracking defects in a bug tracker like Bugzilla, and having an army of human testers. Bringing test assets under source control is beneficial for continuous integration servers, normalization of the QA process, and revision control. A place to run tests should be visible and accessible, such as Sauce Labs. A mechanism to trigger tests should be automated through a continuous integration server to seamlessly integrate with development processes. Logging and tracking defects in a bug tracker that ties tickets to checkins is crucial for visibility and accountability. Human testers are essential for testing what cannot be automated, and their work can be categorized into simple automation, outsourcing intermediate tasks, and utilizing domain experts. The endgame involves planning for human-based QA work and building toward it, which can be achieved by using tools like Sauce Labs and outsourcing testing services.

You can use Behat with Selenium and Sauce to test mobile devices in the cloud. To start, create a Sauce Labs account and obtain your personal username and API key. Configure your Behat tests to use Sauce Labs by installing the Mink Extension and specifying the desired capabilities for your browser, OS, and device. You can then run your tests on iOS and Android devices using the `behat -p` command, and generate HTML reports to visualize your test results. Additionally, you can use Apache ANT to run all your feature files in one go. The source code and usage instructions are available on GitHub. With Behat, Mink Extension, and Sauce Labs, mobile test automation is becoming more accessible and simpler.