go-testing-interface is a Go library that exports an interface that
*testing.T
implements as well as a runtime version you can use in its
place.
The purpose of this library is so that you can export test helpers as a
public API without depending on the "testing" package, since you can't
create a *testing.T
struct manually. This lets you, for example, use the
public testing APIs to generate mock data at runtime, rather than just at
test time.
For usage and examples see the Godoc.
Given a test helper written using go-testing-interface
like this:
import "github.com/mitchellh/go-testing-interface"
func TestHelper(t testing.T) {
t.Fatal("I failed")
}
You can call the test helper in a real test easily:
import "testing"
func TestThing(t *testing.T) {
TestHelper(t)
}
You can also call the test helper at runtime if needed:
import "github.com/mitchellh/go-testing-interface"
func main() {
TestHelper(&testing.RuntimeT{})
}
The tagged version matches the version of Go that the interface is
compatible with. For example, the version "1.14.0" is for Go 1.14 and
introduced the Cleanup
function. The patch version (the ".0" in the
prior example) is used to fix any bugs found in this library and has no
correlation to the supported Go version.
*Why would I call a test helper that takes a testing.T at runtime?
You probably shouldn't. The only use case I've seen (and I've had) for this is to implement a "dev mode" for a service where the test helpers are used to populate mock data, create a mock DB, perhaps run service dependencies in-memory, etc.
Outside of a "dev mode", I've never seen a use case for this and I think
there shouldn't be one since the point of the testing.T
interface is that
you can fail immediately.