Ive heard of stories where people would have an imposed test coverage percentage requirement… and they would just have a single dummy method that printed “.” to the console thousands of times. They then have a single test for that one method, and whenever their codebase grows to big, they add more lines to it so that the dummy method has enough lines to meet the test coverage requirement.
Ive heard of stories where people would have an imposed test coverage percentage requirement… and they would just have a single dummy method that printed “.” to the console thousands of times. They then have a single test for that one method, and whenever their codebase grows to big, they add more lines to it so that the dummy method has enough lines to meet the test coverage requirement.
Percentage of lines covered by tests is such a terrible metric.
All metrics are terrible when used for anything other than objective analysis
Yeah, every metric that can be applied automatically ends up becoming useless if people are incentivized to break it
Relevant xkcd