golang-testing

Define a slice of anonymous structs containing test inputs and expected outputs, then loop over them calling t.Run for each case. This pattern gives complete coverage with minimal code and is the standard Go testing idiom.

Write a function starting with Fuzz that takes *testing.F, add seed corpus entries with f.Add, then define the fuzz function with f.Fuzz. Run it with go test -fuzz=FuzzName -fuzztime=30s; fuzzing requires Go 1.18 or later.

Define an interface for the dependency, then create a mock struct with function fields matching the interface methods. Inject the mock into the code under test, letting you control return values and errors per test case.

Run go test -coverprofile=coverage.out ./... to generate a coverage profile, then use go tool cover -func=coverage.out for per-function stats or go tool cover -html=coverage.out for a browser view. Combine with -race for race detection.

t.Helper() marks a function as a test helper so failure messages report the caller's line number instead of the helper's internal line. This makes debugging failing tests much faster when using shared setup or assertion functions.

Avoid t.Parallel() when subtests share mutable state like a single database connection or depend on execution order. Capture the loop variable before calling t.Parallel() to prevent all subtests from seeing the final iteration value.