I trying to create API with ktor.
Test for ktor api using JUnit5 and I need use the custom properties for database information in 'Application.conf' like below:
ktor {
deployment {
.....
}
application {
.....
}
database {
host = ${?DB_HOST}
user = ${?DB_USER}
pass = ${?DB_PASS}
}
}
I following official guidline for useing custom properties:
HoconApplicationConfig
private val testEnv = createTestEnvironment {
config = HoconApplicationConfig(ConfigFactory.load("application.conf"))
}
class MyTest {
#Test
fun Testing() {
withApplication(testEnv) {
handleRequest(HttpMethod.Get, "/foo").apply {
...
}
}
}
}
Succeed test when only one function in class, but defined second test function then test failed with Exception.
I want define multiple function to simplify the definition and checking in case of failure.
class MyTest {
#Test
fun Testing() {
withApplication(testEnv) {
handleRequest(HttpMethod.Get, "/foo").apply {
....
}
}
}
#Test
fun Testing2() {
withApplication(testEnv) {
handleRequest(HttpMethod.Get, "/foo/bar").apply {
...
}
}
}
}
io.ktor.server.engine.BaseApplicationResponse$ResponseAlreadySentException: Response has already been sent
I know all test will success using 'withTestApplication' but can't use the custom properties.
How can I use the custom properties, and define multiple test funcitons?
Environment:
macOS Big Sur 11.5
IntelliJ Ultimate 2021.1.3
kotlin 1.5.21
ktor 1.6.1
junit-jupiter 5.7.0
Sorry for bad English.
Best Regards.
Related
Update While searching for the optimal solution, I switched temporarily to an approach using a dedicated test factory that performs the init in combination with forAll. In this way, I don't have to repeat the init code in every test factory.
fun somePreStuffFactory() = funSpec {
beforeTest {
println("Do some init stuff here")
}
test("dummy test just to trigger beforeTest") {}
}
class MyTest : FunSpec({
runBlocking {
forAll(
row(MyTestFactory::someTest1),
row(MyTestFactory::someTest2)
) { testFactoryFunc ->
include(somePreStuffFactory())
include(testFactoryFunc())
}
}
})
End of update
I have a lot of tests inside test factories. Before running the tests inside a test factory, I need to do some initial setup in the beginning of each test factory, and the setup is the same for all test factories.
Problem is that beforeSpec is not invoked inside test factory, hence I'm currently using a dirty workaround by doing init stuff in the first test in each test factory. I would highly appreciate any advice on this.
The optimal solution would be to have a life-cycle hook inside the the test class that runs before each test factory.
Code to reproduce
class MyTest : FunSpec({
include(someTest1())
include(someTest2())
})
import io.kotest.core.spec.style.funSpec
object MyTestFactory {
fun someTest1() = funSpec {
beforeSpec {
/** Not invoked */
println("Hello from someTest1#beforeSpec")
}
test("Init stuff is done inside a test") {
/** some init here */
}
test("first test") {
println("Hello from first test")
}
test("second test") {
println("Hello from second test")
}
}
fun someTest2() = funSpec {
beforeSpec {
/** Not invoked */
println("Hello from someTest2#beforeSpec")
}
test("Init stuff is done inside a test") {
/** some init here */
}
test("third test") {
println("Hello from third test")
}
}
}
What I've tried so far
After bumping Kotest from 4.6.4 to 5.4.2, I was able to run the code in answer from #ocos. Problem is that BeforeSpecSample#beforeSpec is invoked just once, not for each test factory which is my requirement.
object BeforeSpecSample : BeforeSpecListener {
override suspend fun beforeSpec(spec: Spec) {
println("Hello from beforeSpec")
}
}
class MyTest : FunSpec({
extensions(BeforeSpecSample)
include(someTest1())
include(someTest2())
})
Update
After reading this GitHub issue, I successfully tested the following approach using a boolean var initialized andbeforeTest. It would have been nice if this approach could be used inside test class instead of inside each test factory, but without any life-cycle hooks for test factories, I don't see how that can be done.
fun someTest1() = funSpec {
var initialized = false
beforeTest {
if (!initialized) {
println("Hello from someTest1#beforeTest")
initialized = true
}
}
/** tests goes here */
}
Environment
Kotest 4.6.4, Kotlin 1.7.10, Micronaut 3.6.3
Kotest test factory doc
If you don't mind grouping everything inside a test factory in a context block, you could use a BeforeContainerListener for initialization:
object InitExtension : BeforeContainerListener {
override suspend fun beforeContainer(testCase: TestCase) {
if(testCase.parent == null) {
println("init stuff")
}
}
}
The if(testCase.parent == null) is just there to allow your factory to have other nested contexts that will not trigger an additional invocation of the initialization.
Then you can write your factory like this:
fun someTest1() = funSpec {
extension(InitExtension)
context("someTest1") {
test("first test") {
println("Hello from first test")
}
test("second test") {
println("Hello from second test")
}
}
}
The initialization in InitExtension will be called at the beginning of context someTest1, and can analogously be included in other test factories that can each be initialized by the same extension.
include is using only tests and extensions defined in your someTest()'s TestFactory. This is why beforeSpec not invoked.
You can register an extension in your MyTest. Here is an example.
// you may use object MyTestFactory : BeforeSpecListener
object BeforeSpecSample : io.kotest.core.listeners.BeforeSpecListener {
override suspend fun beforeSpec(spec: Spec) {
println("-> before spec <-")
}
}
class MyTest : FunSpec({
extensions(BeforeSpecSample)
include(someTest())
test("my test 1") { println("my test 1") }
})
I am a complete beginner in terms of Kotlin and I am finding some issues while trying to test out a Ktor based application.
I have a file in my endpoints package localized at org.example.endpoints.hello
this file contains a fun Application.hello that implements an endpoint for my application.
This endpoint acts as a wrapper for another API, so inside that same file I have a
fun callOtherAPI(): ResponseContainer {
// networking stuff
return ResponseContainer(message: "some stuff")
}
This function gets called inside the Application's function routing implementation as such:
routing {
get("/hello") {
call.respond(callOtherAPI())
}
}
Now to the issue:
My test currently looks like this:
#Test
fun testHello() = testApplication {
application {
hello()
}
mockkStatic(::callOtherAPI)
every { callOtherAPI() } returns ResponseContainer("hello")
print(callOtherAPI()) // This actually returns the mocked response, which is what I want
client.get("/hello").apply {
val expected = ResponseContainer("hello")
val response = jacksonObjectMapper().readValue<ResponseContainer>(bodyAsText())
assertEquals(HttpStatusCode.OK, status)
assertEquals(expected.message, response.message) // This assert fails because the internal call to callOtherAPI() is not being mocked.
}
}
So the problem that I am facing is that while the mocked function is being mocked within the context of the test, it is not being mocked when called internally by the routing implementation.
Can someone point me to good documentation to figure this out, I've been at it for the past two hours to no avail :/
Thanks!
You can declare a parameter for the callOtherAPI function in the hello method. For the production and testing environment you will pass different functions in this case. Here is your code rewritten:
#Test
fun testHello() = testApplication {
application {
// hello(::callOtherAPI) this call will be for the production environment
hello { ResponseContainer("hello") }
}
client.get("/hello").apply {
assertEquals(HttpStatusCode.OK, status)
assertEquals("{\"message\":\"hello\"}", bodyAsText())
}
}
data class ResponseContainer(val message: String)
fun Application.hello(callOtherAPI: () -> ResponseContainer) {
install(ContentNegotiation) {
jackson()
}
routing {
get("/hello") {
call.respond(callOtherAPI())
}
}
}
fun callOtherAPI(): ResponseContainer {
// networking stuff
return ResponseContainer("some stuff")
}
My application (typical REST server that calls other REST services internally) has two main classes to perform the bootstrapping procedure.
There is the Application.kt class that is supposed to configure the vertx instance itself and to register certain modules (jackson kotlin integration for example):
class Application(
private val profileSetting: String? = System.getenv("ACTIVE_PROFILES"),
private val logger: Logger = LoggerFactory.getLogger(Application::class.java)!!
) {
fun bootstrap() {
val profiles = activeProfiles()
val meterRegistry = configureMeters()
val vertx = bootstrapVertx(meterRegistry)
vertx.deployVerticle(ApplicationBootstrapVerticle(profiles)) { startup ->
if (startup.succeeded()) {
logger.info("Application startup finished")
} else {
logger.error("Application startup failed", startup.cause())
vertx.close()
}
}
}
}
In addition there is a ApplicationBootstrapVerticle.kt class that is supposed to deploy the different verticles in a defined order. Some of them in sequence, some of them in parallel:
class ApplicationBootstrapVerticle(
private val profiles: List<String>,
private val logger: Logger = LoggerFactory.getLogger(ApplicationBootstrapVerticle::class.java)
) : AbstractVerticle() {
override fun start(startFuture: Future<Void>) {
initializeApplicationConfig().compose {
logger.info("Application configuration initialized")
initializeScheduledJobs()
}.compose {
logger.info("Scheduled jobs initialized")
initializeRestEndpoints()
}.compose {
logger.info("Http server started")
startFuture
}.setHandler { ar ->
if (ar.succeeded()) {
startFuture.complete()
} else {
startFuture.fail(ar.cause())
}
}
}
private fun initializeApplicationConfig(): Future<String> {
return Future.future<String>().also {
vertx.deployVerticle(
ApplicationConfigVerticle(profiles),
it.completer()
)
}
}
private fun initializeScheduledJobs(): CompositeFuture {
val stationsJob = Future.future<String>()
val capabilitiesJob = Future.future<String>()
return CompositeFuture.all(stationsJob, capabilitiesJob).also {
vertx.deployVerticle(
StationQualitiesVerticle(),
stationsJob.completer()
)
vertx.deployVerticle(
VideoCapabilitiesVerticle(),
capabilitiesJob.completer()
)
}
}
private fun initializeRestEndpoints(): Future<String> {
return Future.future<String>().also {
vertx.deployVerticle(
RestEndpointVerticle(dispatcherFactory = RouteDispatcherFactory(vertx)),
it.completer()
)
}
}
}
I am not sure if this is the supposed way to bootstrap an application, if there is any. More important though, I am not sure if I understand the Future.compose mechanics correctly.
The application starts up successfully and I see all desired log messages except the
Application startup finished
message. Also the following code is never called in case of successs:
}.setHandler { ar ->
if (ar.succeeded()) {
startFuture.complete()
} else {
startFuture.fail(ar.cause())
}
}
In case of an failure though, for example when my application configuration files (yaml) cannot be parsed because there is an unknown field in the destination entity, the log message
Application startup failed
appears in the logs and also the code above is invoked.
I am curious what is wrong with my composed futures chain. I thought that the handler would be called after the previous futures succeeded or one of them failed but I think it's only called in case of success.
Update
I suppose that an invocation of startFuture.complete() was missing. By adapting the start method, it finally worked:
override fun start(startFuture: Future<Void>) {
initializeApplicationConfig().compose {
logger.info("Application configuration initialized")
initializeScheduledJobs()
}.compose {
logger.info("Scheduled jobs initialized")
initializeRestEndpoints()
}.compose {
logger.info("Http server started")
startFuture.complete()
startFuture
}.setHandler(
startFuture.completer()
)
}
I am not sure though, if this is the supposed way to handle this future chain.
The solution that worked for me looks like this:
override fun start(startFuture: Future<Void>) {
initializeApplicationConfig().compose {
logger.info("Application configuration initialized")
initializeScheduledJobs()
}.compose {
logger.info("Scheduled jobs initialized")
initializeRestEndpoints()
}.setHandler { ar ->
if(ar.succeeded()) {
logger.info("Http server started")
startFuture.complete()
} else {
startFuture.fail(ar.cause())
}
}
}
I am using Webflux in Spring Boot 2.0.3.RELEASE to create REST API. With that implementation, I customize and use the webSessionManager as below.
#EnableWebFluxSecurity
#Configuration
class SecurityConfiguration {
#Bean
fun webSessionManager(): WebSessionManager {
return DefaultWebSessionManager().apply {
sessionIdResolver = HeaderWebSessionIdResolver().apply {
headerName = "X-Sample"
}
sessionStore = InMemoryWebSessionStore()
}
}
// ...
}
And in order to test the REST API, I created a test code as follows. (addUser and signin are extension functions.)
#RunWith(SpringRunner::class)
#SpringBootTest
#AutoConfigureWebTestClient
#FixMethodOrder(MethodSorters.NAME_ASCENDING)
class UserTests {
#Autowired
private lateinit var client: WebTestClient
#Test
fun testGetUserInfo() {
client.addUser(defaultUser)
val sessionKey = client.signin(defaultUser)
client.get().uri(userPath)
.header("X-Sample", sessionKey)
.exchange()
.expectStatus().isOk
.expectBody()
.jsonInStrict("""
{
"user": {
"mail_address": "user#example.com"
}
}
""".trimIndent())
}
// ...
}
The test failed. It is refused by authorization. However, if I start the server and run it from curl it will succeed in the authorization.
After investigating the cause, it turned out that org.springframework.test.web.reactive.server.AbstractMockServerSpec set webSessionManager to DefaultWebSessionManager. Default is used, not the webSessionManager I customized. For this reason, it could not get the session ID.
AbstractMockServerSpec.java#L41
AbstractMockServerSpec.java#L72-L78
How can I change the webSessionManager of AbstractMockServerSpec?
Also, I think that it is better to have the following implementation, what do you think?
abstract class AbstractMockServerSpec<B extends WebTestClient.MockServerSpec<B>>
implements WebTestClient.MockServerSpec<B> {
// ...
private WebSessionManager sessionManager = null;
// ...
#Override
public WebTestClient.Builder configureClient() {
WebHttpHandlerBuilder builder = initHttpHandlerBuilder();
builder.filters(theFilters -> theFilters.addAll(0, this.filters));
if (this.sessionManager != null) {
builder.sessionManager(this.sessionManager);
}
this.configurers.forEach(configurer -> configurer.beforeServerCreated(builder));
return new DefaultWebTestClientBuilder(builder);
}
// ...
}
Spring Framework's AbstractMockServerSpec is providing a method to customize the WebSessionManager already.
Thanks for opening SPR-17094, this problem will be solved with that ticket - the AbstractMockServerSpec is already looking into the application context for infrastructure bits, this should check for a WebSessionManager as well.
I am trying to test my AkkaHTTP routes (written in Kotlin) using akka-http-testkit. The tests in our project use Spek and I would like to keep it this way.
The Route TestKit tutorial gives a Java example:
public class TestkitExampleTest extends JUnitRouteTest {
TestRoute appRoute = testRoute(new MyAppService().createRoute())
#Test
public void testCalculatorAdd() {
// test happy path
appRoute.run(HttpRequest.GET("/calculator/add?x=4.2&y=2.3"))
.assertStatusCode(200)
.assertEntity("x + y = 6.5")
// test responses to potential errors
appRoute.run(HttpRequest.GET("/calculator/add?x=3.2"))
.assertStatusCode(StatusCodes.NOT_FOUND) // 404
.assertEntity("Request is missing required query parameter 'y'")
// test responses to potential errors
appRoute.run(HttpRequest.GET("/calculator/add?x=3.2&y=three"))
.assertStatusCode(StatusCodes.BAD_REQUEST)
.assertEntity("The query parameter 'y' was malformed:\n" +
"'three' is not a valid 64-bit floating point value")
}
}
The setup uses the testRoute function, which means the test class must extend JUnitRouteTest.
Attempting to translate to a Kotlin Spek test I got this:
class TestKitExampleTest : JUnitRouteTest(), Spek({
describe("My routes") {
val appRoute = testRoute(MyAppService().createRoute())
it("calculator add") {
// test happy path
appRoute.run(HttpRequest.GET("/calculator/add?x=4.2&y=2.3"))
.assertStatusCode(200)
.assertEntity("x + y = 6.5")
//...rest omitted
}
}
})
which does not compile as the class is attempting to inherit two classes. I converted it to the following instead:
class TestKitExampleTest : Spek({
describe("My routes") {
val appRoute = testRoute(MyAppService().createRoute())
it("calculator add") {
// test happy path
appRoute.run(HttpRequest.GET("/calculator/add?x=4.2&y=2.3"))
.assertStatusCode(200)
.assertEntity("x + y = 6.5")
//...rest omitted
}
}
}) {
companion object : JUnitRouteTest()
}
which encouters the runtime error java.lang.IllegalStateException: Unknown factory null
at akka.http.impl.util.package$.actorSystem(package.scala:34).
Is there a way to use Akka's route testkit with Spek? Or is there another way to test these routes?
As #raniejade mentioned above, answered on Github. JUnitRouteTest bootstraps Akka with a rule, but Spek's LifeCycleListener can do the same thing.
Adding the code:
class SpekRouteBootstrapper: LifecycleListener, JUnitRouteTest() {
override fun beforeExecuteTest(test: TestScope) {
systemResource().before()
}
override fun afterExecuteTest(test: TestScope) {
systemResource().after()
}
}
allowed me to do this on the test class:
class TestKitExampleTest: Spek({
val bootstrapper = SpekRouteBootstrapper()
registerListener(bootstrapper)
describe("My routes") {
val appRoute by memoized {
bootstrapper.testRoute(MyAppService().createRoute())
}
it("calculator add") {
// test happy path
appRoute.run(HttpRequest.GET("/calculator/add?x=4.2&y=2.3"))
.assertStatusCode(200)
.assertEntity("x + y = 6.5")
}
}
})