add an option to start the local debugging server based on an env var…

…iable (#87)

motivation: make using the local debugging server easier to turn off/on without the need to change code when oyu are preparing to deploy

changes:
* add code to lambda so that in debug mode only, if the LOCAL_LAMBDA_SERVER_ENABLED env variable is set the local debugging server is started
* make withLocalServer internal
* update example code

Examples/LocalDebugging/MyLambda/Sources/MyLambda/main.swift

@@ -15,9 +15,9 @@
 import AWSLambdaRuntime
 import Shared
 
-try Lambda.withLocalServer {
- Lambda.run { (_: Lambda.Context, request: Request, callback: @escaping (Result<Response, Error>) -> Void) in
- // TODO: something useful
-  callback(.success(Response(message: "Hello, \(request.name)!")))
- }
+// set LOCAL_LAMBDA_SERVER_ENABLED env variable to "true" to start
+// a local server simulator which will allow local debugging
+Lambda.run { (_, request: Request, callback: @escaping (Result<Response, Error>) -> Void) in
+ // TODO: something useful
+ callback(.success(Response(message: "Hello, \(request.name)!")))
 }

Examples/LocalDebugging/README.md

@@ -11,18 +11,25 @@ The example includes three modules:
 3. [Shared](Shared) is a SwiftPM library package used for shared code between the iOS application and the Lambda function,
 such as the Request and Response model objects.
 
-The local debugging experience is achieved by running the Lambda function in the context of the debug only `Lambda.withLocalServer`
-function which starts a local emulator enabling the communication
+The local debugging experience is achieved by running the Lambda function in the context of the
+debug-only local lambda engine simulator which starts a local HTTP server enabling the communication
 between the iOS application and the Lambda function over HTTP.
 
 To try out this example, open the workspace in Xcode and "run" the two targets,
 using the relevant `MyLambda` and `MyApp` Xcode schemas.
 
-Start with running the `MyLambda` target on the "My Mac" destination, once it is up you should see a log message in the Xcode console saying
+Start with running the `MyLambda` target.
+* Switch to the `MyApp` scheme and select the "My Mac" destination
+* Set the `LOCAL_LAMBDA_SERVER_ENABLED` environment variable to `true` by editing the `MyLambda` scheme under `Run`.
+* Hit `Run`
+* Once it is up you should see a log message in the Xcode console saying
 `LocalLambdaServer started and listening on 127.0.0.1:7000, receiving payloads on /invoke`
 which means the local emulator is up and receiving traffic on port `7000` and expecting payloads on the `/invoke` endpoint.
 
-Continue to run the `MyApp` target in a simulator destination. Once up, the application's UI should appear in the simulator allowing you
+Continue to run the `MyApp` target
+* Switch to the `MyApp` scheme and select a simulator destination.
+* Hit `Run`
+* Once up, the application's UI should appear in the simulator allowing you
 to interact with it.
 
 Once both targets are running, set up breakpoints in the iOS application or Lambda function to observe the system behavior.

Sources/AWSLambdaRuntimeCore/Lambda+LocalServer.swift

@@ -35,11 +35,12 @@ extension Lambda {
  /// - body: Code to run within the context of the mock server. Typically this would be a Lambda.run function call.
  ///
  /// - note: This API is designed stricly for local testing and is behind a DEBUG flag
- public static func withLocalServer(invocationEndpoint: String? = nil, _ body: @escaping () -> Void) throws {
+ @discardableResult
+ static func withLocalServer<Value>(invocationEndpoint: String? = nil, _ body: @escaping () -> Value) throws -> Value {
  let server = LocalLambda.Server(invocationEndpoint: invocationEndpoint)
  try server.start().wait()
  defer { try! server.stop() } // FIXME:
- body()
+ return body()
  }
 }
 

Sources/AWSLambdaRuntimeCore/Lambda.swift

@@ -97,36 +97,55 @@ public enum Lambda {
  // for testing and internal use
  @discardableResult
  internal static func run(configuration: Configuration = .init(), factory: @escaping HandlerFactory) -> Result<Int, Error> {
- Backtrace.install()
- var logger = Logger(label: "Lambda")
- logger.logLevel = configuration.general.logLevel
+ let _run = { (configuration: Configuration, factory: @escaping HandlerFactory) -> Result<Int, Error> in
+ Backtrace.install()
+ var logger = Logger(label: "Lambda")
+ logger.logLevel = configuration.general.logLevel
 
- var result: Result<Int, Error>!
- MultiThreadedEventLoopGroup.withCurrentThreadAsEventLoop { eventLoop in
- let lifecycle = Lifecycle(eventLoop: eventLoop, logger: logger, configuration: configuration, factory: factory)
- #if DEBUG
- let signalSource = trap(signal: configuration.lifecycle.stopSignal) { signal in
- logger.info("intercepted signal: \(signal)")
- lifecycle.shutdown()
- }
- #endif
-
- lifecycle.start().flatMap {
- lifecycle.shutdownFuture
- }.whenComplete { lifecycleResult in
+ var result: Result<Int, Error>!
+ MultiThreadedEventLoopGroup.withCurrentThreadAsEventLoop { eventLoop in
+ let lifecycle = Lifecycle(eventLoop: eventLoop, logger: logger, configuration: configuration, factory: factory)
  #if DEBUG
- signalSource.cancel()
+ let signalSource = trap(signal: configuration.lifecycle.stopSignal) { signal in
+ logger.info("intercepted signal: \(signal)")
+ lifecycle.shutdown()
+ }
  #endif
- eventLoop.shutdownGracefully { error in
- if let error = error {
- preconditionFailure("Failed to shutdown eventloop: \(error)")
+
+ lifecycle.start().flatMap {
+ lifecycle.shutdownFuture
+ }.whenComplete { lifecycleResult in
+ #if DEBUG
+ signalSource.cancel()
+ #endif
+ eventLoop.shutdownGracefully { error in
+ if let error = error {
+ preconditionFailure("Failed to shutdown eventloop: \(error)")
+ }
  }
+ result = lifecycleResult
  }
- result = lifecycleResult
  }
+
+ logger.info("shutdown completed")
+ return result
  }
 
- logger.info("shutdown completed")
- return result
+ // start local server for debugging in DEBUG mode only
+ #if DEBUG
+ if Lambda.env("LOCAL_LAMBDA_SERVER_ENABLED").flatMap(Bool.init) ?? false {
+ do {
+ return try Lambda.withLocalServer {
+ _run(configuration, factory)
+ }
+ } catch {
+ return .failure(error)
+ }
+ } else {
+ return _run(configuration, factory)
+ }
+ #else
+ return _run(configuration, factory)
+ #endif
  }
 }