golang-testing

Define a slice of anonymous structs with name, inputs, and expected outputs, then loop over it calling t.Run(tt.name, ...) for each case. This pattern gives one subtest per case with clear failure output and minimal code duplication.

Write a BenchmarkXxx function taking *testing.B and loop with b.Loop() (Go 1.24+) or for i := 0; i < b.N; i++. Run with go test -bench=. -benchmem to report ns/op, bytes per operation, and allocations per operation.

Yes, Go 1.18+ includes built-in fuzzing via testing.F. Add seed inputs with f.Add(), define the fuzz function with f.Fuzz(), and run it using go test -fuzz=FuzzName -fuzztime=30s to discover edge-case inputs automatically.

t.Fatal and t.Fatalf only terminate the calling goroutine, so inside a handler or spawned goroutine they truncate the response without failing the test properly. Use t.Errorf plus return, or record errors with atomics and assert on the main test goroutine.

Run go test -race -coverprofile=coverage.out ./... then inspect with go tool cover -func=coverage.out. Parse the total percentage and fail the pipeline if it falls below your threshold, such as the 90% gate used in this project.