Stability in Unstability: Self‑Healing Test Automation
summary
Describes a reusable, framework‑agnostic self‑healing layer for test automation that intercepts Selenium element-not-found errors and replaces broken static locators by comparing previous and current DOMs using graph edit distance rather than only visual matching. Prototypes generate new locators from modified nodes and were validated against a random DOM generator; the approach handles attribute tweaks and saved examples reliably but is memory‑hungry and slow on large DOM diffs, and the talk references commercial tools (Mabl, Testim.io, Applitools) and prior projects that experienced about 40% locator failures.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 6:10Testing evolution
- 10:07Self-healing automation
- 14:42Neural networks in testing
- 20:02Stabilizing visual locators
- 27:43Legacy projects approach
- 36:28DOM comparison proof
- 41:55Q&A
- 51:23Summary
related talks
54mStability in Unstability: Self-Healing Test Automation
30m'Self-Healing Tests' - The holy grail of test automation?
25mSelf-Healing in Test Automation
31mSelf-Healing or Self-Killing?
54mSelf-Healing Automation with Healenium and Regression Minimization with Drill4J
43mSmart and Stable Automated Testing Using AI Based Dynamic Locators
32mSelf-healing system for UI tests using ML
44mSelf-healing Test Automation 2.0: The Future