Code Coverage Limitations¶

Code coverage measures which code paths executed during testing, not whether the tests verified correct behavior - high coverage with weak assertions provides false confidence, while meaningful tests might not maximize coverage metrics.

Tension Addressed¶

Coverage vs. Quality

Recognizable As¶

Coverage reports showing 90%+ but bugs still reaching production; tests without assertions; assertions only checking "no exception thrown"; coverage mandates without quality review

Canonical Sources¶

Marick, "How to Misuse Code Coverage" (1999); Fowler, "TestCoverage" (2012)

Transfers To¶

Goodhart's Law (measure becomes target); vanity metrics in analytics; audit compliance vs. actual security

Coverage is a necessary but not sufficient condition for test quality. A line is "covered" if it executes, regardless of whether the test verifies its behavior. The metric is useful for finding untested code but cannot distinguish meaningful tests from trivial ones.