Rule: A failing automated check should break the build
Rule: A failing automated check should break the build.
If we start allowing some warnings, errors, or broken tests, it creates a slippery slope.
I’ve seen teams with 100’s of ignored warnings, and dozens of ignored broken tests. It started with just a few, then they slowly became blind to the growing problem.
Plus, existing warnings and broken tests mask new issues, so new problems just pile up.
The solution: Fail the build if there’s a single warning, error, or broken test. Don’t allow quality to slide backward.