Operwatir and Testing at Opera
summary
Opera uses Selenium plus a native backend to run browser tests at scale with concrete results: about 200,000 tests across ~180 slaves (roughly 160 Linux machines and ~20 Windows), producing 1–6 million results daily. Automated tools include OperaDriver and Watir bindings, RSpec tests, screenshot comparison and C++ self tests. Regression fixes historically introduce ~30% new bugs, so unstable tests enter a quarantine phase before joining stable suites; visual hashes and per-element screenshots cut false positives.
AI-written summary · about
chapters (9)
- 0:00Automated testing at Opera
- 1:13Writing a web browser complexity
- 4:23Regression testing challenges
- 7:37Performance testing in Opera
- 9:09Automated regression success
- 12:18Stability testing false positives
- 15:50Automated test types
- 17:37Selenium for browser testing
- 21:12Standardizing browser automation APIs
related talks
41mTesting The Web Anywhere: OperaDriver on Mobile Devices
30mLessons from testing browsers using WebDriver
37mAutomated Performance Testing With WebDriver
42mReliable and Maintainable Webdriver Tests
19mWorking Well: The Future of Web Testing
37mSelenium In the Enterprise: What went right, what went wrong (so far)
50mTesting in Production – dangerous, scary or better?
30mBrowser Testing