Metrics · Updated 2026-09-17
Code coverage
The percentage of code executed by the test suite — a measure of what is tested, not of how well it is tested.
What it is
Instrument the code, run the tests, record which lines or branches executed. Line coverage is the common headline; branch coverage is stricter and more informative.
Why the number is weaker than it looks
Coverage tells you code ran during a test. It does not tell you the test asserted anything meaningful about it. A suite that executes every line and asserts almost nothing reports high coverage and catches nothing — and this is not hypothetical, it is what happens whenever a coverage threshold becomes a target.
The useful version is directional: coverage on changed code, in a pull request. “This change added forty lines and tested none of them” is actionable. “The repository is at 73%” is not.
Why it appears in review tooling
Several quality platforms post coverage deltas as a pull request check, and it is one of the better automated review signals — objective, relevant to the change at hand, and hard to argue with.
Common mistakes
- Setting a global percentage target, which produces tests written to touch lines rather than to verify behaviour.
- Treating coverage as a proxy for quality in a report to stakeholders.
- Ignoring which code is uncovered. Untested error handling matters more than an untested getter.