Lightning Talk: Cpp Change Detector Tests
summary
change detector tests intentionally fail on any non-trivial code change and are used to surface concrete differences: an AI pet image classifier example shows overall accuracy and per-class regressions, cache hit/miss rates spike unexpectedly, and performance numbers shift. Tests generate a human-readable string of metrics that appears in failures and is trivially updated by copying the error output. This puts soft decisions into code review and preserves historic performance snapshots in the repo for investigation.
AI-written summary · about
related talks
32mLightning Talks
44mTesting The Tests: Mutation Testing for C++
01h 03mRewiring your brain with test driven thinking in C++
01h 40mTesting, Preconditions, Coverage and Templates in Safety-Critical C++ Code
11mUplifting the Art of Code Review from Testing Perspective
05mLightning Talk: Zero Cost Regressions
42mTesting on Thin Ice: Chipping Away at Test Unpredictability
01h 00mContinuous Regression Testing for Safer and Faster Refactoring in C++