Marharyta Nedzelska at JNation 2024
summary
Collects common myths about code coverage and shows concrete pitfalls: JaCoCo reports that mark switch statements as uncovered even when exhaustive, teams chasing 80% coverage target end up adding pointless mocks or testing untestable artifacts, and Sonar quality gates force numbers without a way to mark coverage as accepted or false positive. Explains which bytecode patterns cause false negatives, which engineering hacks worsen production code, and why tools need an “accept coverage” workflow rather than a single percent to gate builds.
AI-written summary · about
related talks
08mInterview of Marharyta Nedzelska
46mJava, Code Coverage & Their Best Friend: Bytecode
38mThe state of code coverage for Kotlin
51mDebunking Myths Around Test Coverage
10m99% is not enough - full test coverage and why it's awesome
58mCoverage, quality, and Goodhart's Law
02h 46mThe Joy of Testing
03mTest coverage