“Save $$ Testing Code the Playback-Based Way”
summary
Advocates recording every API call, arguments and return values into JSON artifacts and replaying them against the library to catch breaking changes; a header-only C++ CppPlayer prototype (≈1,200 lines) supports both recording and playback, does pointer swizzling and compares function returns, and is used to tame a flaky HFDM cloud-storage integration. Proven at scale in CAD: Revit journaling produced ~15,000–19,000 automated tests. Costs include added instrumentation/boilerplate and an API-layer wrapper; it works but requires careful handling of threads and test curation. Playback-Based Testing recordings
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:06Playback-based testing for C++
- 3:40Library scaling dilemma
- 5:10Qualifying library integration
- 7:44Refactoring for playback testing
- 12:41Case study: Revit Journaling
- 15:58Introducing CppPlayer
- 17:44Growing your own tool
- 23:37Implementation details
- 28:25Q&A
related talks
46mFrom Betamax to VCR: harden your API integration testing
46mRecord-Replay API Test Cases & Data Mocks without writing code
27mThe Joy of Record and Playback in Test Automation
51mAdvanced Playwright Techniques for Flawless Testing
14mAdvanced Playwright Techniques for flawless testing
08mHow Big Buck Bunny helped track down bugs
45mTest Automation Requirements Through Recordings
24mFighting Test Flakiness with Time Machines