import snunit.*

@main
def run =
  SyncServerBuilder
    .setRequestHandler(req =>
      req.send(
        statusCode = StatusCode.OK,
        content = "Hello world!\n",
        headers = Headers("Content-Type" -> "text/plain")
      )
    )
    .build()
    .listen()

SNUnit is a Scala Native library to write HTTP server applications on top of NGINX Unit. It allows you to write both synchronous and asynchronous web servers with automatic restart on crashes, automatic load balancing of multiple processes, great performance and all the nice NGINX Unit features.

Once built your SNUnit binary, you need to deploy it to the unitd server.

You need to run unitd in a terminal with:

unitd --no-daemon --log /dev/stdout --control unix:control.sock

This will run unitd with a UNIX socket file named control.sock in your current directory.

Then, you need to create a json file with your configuration:

{
  "listeners": {
    "*:8081": {
      "pass": "applications/myapp"
    }
  },
  "applications": {
    "myapp": {
      "type": "external",
      "executable": "snunit/binary/path"
    }
  }
}

Where executable is the binary path which can be absolute or relative to the unitd working directory.

This configuration passes all requests sent to the port 8081 to the application myapp.

To know more about configuring NGINX Unit, refer to its documentation.

To deploy the setting you can use curl:

curl -X PUT --unix-socket control.sock -d @config.json localhost/config

If everything went right, you should see this response:

{
  "success": "Reconfiguration done."
}

In case of problems, you will get a 4xx response like this:

{
  "error": "Invalid configuration.",
  "detail": "Required parameter \"executable\" is missing."
}

Further information can be found in unitd logs in the running terminal.

SNUnit has two different server implementations.

With SyncServerBuilder you need to call .listen() to start listening. It is a blocking operation so your process is stuck on listening and can't do anything else while listening. Moreover, all the request handlers need to respond directly and can't be implemented using Futures or any other asyncronous mechanism since no Future will run, being the process stuck on the listen() Unit event loop. With http4s or tapir-cats-effect the server is automatically scheduled to run either on the cats effect event loop, based on epoll/kqueue. This allows you to complete requests asyncronously using whatever mechanism you prefer. A process can accept multiple requests concurrently, allowing great parallelism.

SNUnit offers interpreters for Tapir server endpoints. You can write all your application using Tapir and the convert your Tapir endpoints with logic into a SNUnit Handler.

Currently two interpreters are available:

• SNUnitIdServerInterpreter which works best with SyncServerHandler for synchronous applications
  • You can find an example in tests
• An interpreter for cats hidden behind snunit.tapir.SNUnitServerBuilder in the snunit-tapir-cats-effect artifact.
  • You can find an example in tests

snunit.TapirApp extends cats.effect.IOApp building the SNUnit server.

It exposes a def serverEndpoints: Resource[IO, List[ServerEndpoint[Any, IO]]] that you need to implement with your server logic.

Here an example "Hello world" app:

import cats.effect.*
import sttp.tapir.*

object Main extends snunit.TapirApp {
  def serverEndpoints = Resource.pure(
    endpoint.get
      .in("hello")
      .in(query[String]("name"))
      .out(stringBody)
      .serverLogic[IO](name => IO(Right(s"Hello $name!"))) :: Nil
  )
}

SNUnit offers a server implementation for http4s. It is based on the epollcat asynchronous event loop.

There are two ways you can build a http4s server.

snunit.Http4sApp extends cats.effect.IOApp building the SNUnit server.

It exposes a def routes: Resource[IO, HttpApp[IO]] that you need to implement with your server logic.

Here an example "Hello world" app:

import cats.effect.*
import org.http4s.*
import org.http4s.dsl.io.*

object app extends snunit.Http4sApp {
  def routes = Resource.pure(
    HttpRoutes
      .of[IO] { case GET -> Root =>
        Ok("Hello from SNUnit Http4s!")
      }
      .orNotFound
  )
}

If you want to have more control over the server creation, you can use the SNUnitServerBuilder and manually use it.

For example, here you see it in combination with cats.effect.IOApp

package snunit.tests

import cats.effect.*
import org.http4s.*
import org.http4s.dsl.io.*
import snunit.http4s.*

object Http4sHelloWorld extends IOApp.Simple {
  def helloWorldRoutes: HttpRoutes[IO] =
    HttpRoutes.of[IO] { case GET -> Root =>
      Ok("Hello Http4s!")
    }

  def run: IO[Unit] =
    SNUnitServerBuilder
      .default[IO]
      .withHttpApp(helloWorldRoutes.orNotFound)
      .run
}