skinny.

Latest / Software Testing with Fexingo: QA, Automation, and Reliable Software Engineering

Why Test Coverage Metrics Lie About Code Quality

Lucas and Luna dive into a dangerous habit in software quality: chasing high test coverage percentages. Lucas explains why code coverage numbers often mask untested logic paths, citing a 2024 study showing that 40% of 'covered' lines in real-world JavaScript projects had zero assertions. They discuss structural versus decision coverage, the trap of mocking everything, and why a 70% line coverage with strong assertions beats 95% with shallow tests. Luna brings in a case from a fintech startup where high coverage gave false confidence before a production bug wiped out a payment flow…

The skinny

The skinny isn't ready yet — notes appear once the transcript is processed.