How does the testing package support table-driven tests and benchmarks?

Table-driven tests use a slice of test cases iterated with t.Run() for isolated subtests. Benchmarks use the testing.B type with a b.N loop that the runtime calibrates automatically.

Table-driven tests and benchmarks

Testing best practices

1
t.Run() creates isolated subtests — failures are reported individually and can be run in parallel with t.Parallel()
2
Use testify/assert for cleaner assertion messages: assert.Equal(t, expected, actual)
3
b.ResetTimer() to exclude setup time from benchmark measurements
4
go test -race -count=10 catches non-deterministic race conditions
5
Examples (func ExampleXxx) serve as documentation and are verified by go test