concurrent

Type Members

1. trait AbstractPatienceConfiguration extends ScaledTimeSpans

   

   Trait that defines an abstract patienceConfig method that is implemented in PatienceConfiguration and can be overriden in stackable modification traits such as IntegrationPatience.

   Trait that defines an abstract patienceConfig method that is implemented in PatienceConfiguration and can be overriden in stackable modification traits such as IntegrationPatience.

   The main purpose of AbstractPatienceConfiguration is to differentiate core PatienceConfiguration traits, such as Eventually and AsyncAssertions, from stackable modification traits for PatienceConfigurations such as IntegrationPatience. Because these stackable traits extend AbstractPatienceConfiguration instead of Suite, you can't simply mix in a stackable trait:

   class ExampleSpec extends FunSpec with IntegrationPatience // Won't compile
   

   The previous code is undesirable because IntegrationPatience would have no affect on the class. Instead, you need to mix in a core PatienceConfiguration trait and mix the stackable IntegrationPatience trait into that, like this:

   class ExampleSpec extends FunSpec with Eventually with IntegrationPatience // Compiles fine
   

   The previous code is better because IntegrationPatience does have an effect: it modifies the behavior of Eventually.

2. trait AsyncAssertions extends PatienceConfiguration

   

   Trait that facilitates performing assertions outside the main test thread, such as assertions in callback methods that are invoked asynchronously.

   Trait that facilitates performing assertions outside the main test thread, such as assertions in callback methods that are invoked asynchronously.

   Trait AsyncAssertions provides a Waiter class that you can use to orchestrate the inter-thread communication required to perform assertions outside the main test thread, and a means to configure it.

   To use Waiter, create an instance of it in the main test thread:

   val w = new Waiter // Do this in the main test thread
   

   At some point later, call await on the waiter:

   w.await() // Call await() from the main test thread
   

   The await call will block until it either receives a report of a failed assertion from a different thread, at which point it will complete abruptly with the same exception, or until it is dismissed by a different thread (or threads), at which point it will return normally. You can optionally specify a timeout and/or a number of dismissals to wait for. Here's an example:

   import org.scalatest.time.SpanSugar._
   
   w.await(timeout(300 millis), dismissals(2))
   

   The default value for timeout, provided via an implicit PatienceConfig parameter, is 150 milliseconds. The default value for dismissals is 1. The await method will block until either it is dismissed a sufficient number of times by other threads or an assertion fails in another thread. Thus if you just want to perform assertions in just one other thread, only that thread will be performing a dismissal, so you can use the default value of 1 for dismissals.

   Waiter contains four overloaded forms of await, two of which take an implicit PatienceConfig parameter. To change the default timeout configuration, override or hide (if you imported the members of AsyncAssertions companion object instead of mixing in the trait) patienceConfig with a new one that returns your desired configuration.

   To dismiss a waiter, you just invoke dismiss on it:

   w.dismiss() // Call this from one or more other threads
   

   You may want to put dismiss invocations in a finally clause to ensure they happen even if an exception is thrown. Otherwise if a dismissal is missed because of a thrown exception, an await call will wait until it times out.

   Note that if a Waiter receives more than the expected number of dismissals, it will not report this as an error: i.e., receiving greater than the number of expected dismissals without any failed assertion will simply cause the the test to complete, not to fail. The only way a Waiter will cause a test to fail is if one of the asynchronous assertions to which it is applied fails.

   Finally, to perform an assertion in a different thread, you just apply the Waiter to the assertion code. Here are some examples:

   w { assert(1 + 1 === 3) }    // Can use assertions
   w { 1 + 1 should equal (3) } // Or matchers
   w { "hi".charAt(-1) }        // Any exceptions will be forwarded to await
   

   Here's a complete example:

   import org.scalatest._
   import concurrent.AsyncAssertions
   import scala.actors.Actor
   
   class ExampleSuite extends FunSuite with Matchers with AsyncAssertions {
   
     case class Message(text: String)
   
     class Publisher extends Actor {
   
       @volatile private var handle: Message => Unit = { (msg) => }
   
       def registerHandler(f: Message => Unit) {
         handle = f
       }
   
       def act() {
         var done = false
         while (!done) {
           react {
             case msg: Message => handle(msg)
             case "Exit" => done = true
           }
         }
       }
     }
   
     test("example one") {
   
       val publisher = new Publisher
       val message = new Message("hi")
       val w = new Waiter
   
       publisher.start()
   
       publisher.registerHandler { msg =>
         w { msg should equal (message) }
         w.dismiss()
       }
   
       publisher ! message
       w.await()
       publisher ! "Exit"
     }
   }
   

3. trait ConductorFixture extends SuiteMixin with Conductors

   

   Trait that can pass a new Conductor fixture into tests.

   Trait that can pass a new Conductor fixture into tests.

   Here's an example of the use of this trait to test the ArrayBlockingQueue class from java.util.concurrent:

   import org.scalatest.fixture
   import org.scalatest.concurrent.ConductorFixture
   import org.scalatest.matchers.ShouldMatchers
   import java.util.concurrent.ArrayBlockingQueue
   
   class ArrayBlockingQueueSuite extends fixture.FunSuite with ConductorFixture with ShouldMatchers {
   
     test("calling put on a full queue blocks the producer thread") { conductor => import conductor._
   
       val buf = new ArrayBlockingQueue[Int](1)
   
       thread("producer") {
         buf put 42
         buf put 17
         beat should be (1)
       }
   
       thread("consumer") {
         waitForBeat(1)
         buf.take should be (42)
         buf.take should be (17)
       }
   
       whenFinished {
         buf should be ('empty)
       }
     }
   
     test("calling take on an empty queue blocks the consumer thread") { conductor => import conductor._
   
       val buf = new ArrayBlockingQueue[Int](1)
   
       thread("producer") {
         waitForBeat(1)
         buf put 42
         buf put 17
       }
   
       thread("consumer") {
         buf.take should be (42)
         buf.take should be (17)
         beat should be (1)
       }
   
       whenFinished {
         buf should be ('empty)
       }
     }
   }
   

   For an explanation of how these tests work, see the documentation for Conductors.

4. trait ConductorMethods extends SuiteMixin with Conductors

   

   Trait that provides each test with access to a new Conductor via methods.

   Trait that provides each test with access to a new Conductor via methods.

   Here's an example of the use of this trait to test the ArrayBlockingQueue concurrency abstraction from java.util.concurrent:

   import org.scalatest.FunSuite
   import org.scalatest.concurrent.ConductorMethods
   import org.scalatest.matchers.ShouldMatchers
   import java.util.concurrent.ArrayBlockingQueue
   
   class ArrayBlockingQueueSuite extends FunSuite with ConductorMethods with ShouldMatchers {
   
     test("calling put on a full queue blocks the producer thread") {
   
       val buf = new ArrayBlockingQueue[Int](1)
   
       thread("producer") {
         buf put 42
         buf put 17
         beat should be (1)
       }
   
       thread("consumer") {
         waitForBeat(1)
         buf.take should be (42)
         buf.take should be (17)
       }
   
       whenFinished {
         buf should be ('empty)
       }
     }
   
     test("calling take on an empty queue blocks the consumer thread") {
   
       val buf = new ArrayBlockingQueue[Int](1)
   
       thread("producer") {
         waitForBeat(1)
         buf put 42
         buf put 17
       }
   
       thread("consumer") {
         buf.take should be (42)
         buf.take should be (17)
         beat should be (1)
       }
   
       whenFinished {
         buf should be ('empty)
       }
     }
   }
   

   For an explanation of how these tests work, see the documentation for Conductors.

5. trait Conductors extends PatienceConfiguration

   

   Trait whose Conductor member facilitates the testing of classes, traits, and libraries designed to be used by multiple threads concurrently.

   Trait whose Conductor member facilitates the testing of classes, traits, and libraries designed to be used by multiple threads concurrently.

   A Conductor conducts a multi-threaded scenario by maintaining a clock of "beats." Beats are numbered starting with 0. You can ask a Conductor to run threads that interact with the class, trait, or library (the subject) you want to test. A thread can call the Conductor's waitForBeat method, which will cause the thread to block until that beat has been reached. The Conductor will advance the beat only when all threads participating in the test are blocked. By tying the timing of thread activities to specific beats, you can write tests for concurrent systems that have deterministic interleavings of threads.

   A Conductor object has a three-phase lifecycle. It begins its life in the setup phase. During this phase, you can start threads by invoking the thread method on the Conductor. When conduct is invoked on a Conductor, it enters the conducting phase. During this phase it conducts the one multi-threaded scenario it was designed to conduct. After all participating threads have exited, either by returning normally or throwing an exception, the conduct method will complete, either by returning normally or throwing an exception. As soon as the conduct method completes, the Conductor enters its defunct phase. Once the Conductor has conducted a multi-threaded scenario, it is defunct and can't be reused. To run the same test again, you'll need to create a new instance of Conductor.

   Here's an example of the use of Conductor to test the ArrayBlockingQueue class from java.util.concurrent:

   import org.scalatest.fixture.FunSuite
   import org.scalatest.matchers.ShouldMatchers
   import java.util.concurrent.ArrayBlockingQueue
   import org.scalatest.concurrent.Conductors
   
   class ArrayBlockingQueueSuite extends FunSuite with ShouldMatchers with Conductors {
   
     test("calling put on a full queue blocks the producer thread") {
   
       val conductor = new Conductor
       import conductor._
   
       val buf = new ArrayBlockingQueue[Int](1)
   
       thread("producer") {
         buf put 42
         buf put 17
         beat should be (1)
       }
   
       thread("consumer") {
         waitForBeat(1)
         buf.take should be (42)
         buf.take should be (17)
       }
   
       whenFinished {
         buf should be ('empty)
       }
     }
   }
   

   When the test shown is run, it will create one thread named producer and another named consumer. The producer thread will eventually execute the code passed as a by-name parameter to thread("producer"):

   buf put 42
   buf put 17
   beat should be (1)
   

   Similarly, the consumer thread will eventually execute the code passed as a by-name parameter to thread("consumer"):

   waitForBeat(1)
   buf.take should be (42)
   buf.take should be (17)
   

   The thread creates the threads and starts them, but they will not immediately execute the by-name parameter passed to them. They will first block, waiting for the Conductor to give them a green light to proceed.

   The next call in the test is whenFinished. This method will first call conduct on the Conductor, which will wait until all threads that were created (in this case, producer and consumer) are at the "starting line", i.e., they have all started and are blocked, waiting on the green light. The conduct method will then give these threads the green light and they will all start executing their blocks concurrently.

   When the threads are given the green light, the beat is 0. The first thing the producer thread does is put 42 in into the queue. As the queue is empty at this point, this succeeds. The producer thread next attempts to put a 17 into the queue, but because the queue has size 1, this can't succeed until the consumer thread has read the 42 from the queue. This hasn't happened yet, so producer blocks. Meanwhile, the consumer thread's first act is to call waitForBeat(1). Because the beat starts out at 0, this call will block the consumer thread. As a result, once the producer thread has executed buf put 17 and the consumer thread has executed waitForBeat(1), both threads will be blocked.

   The Conductor maintains a clock that wakes up periodically and checks to see if all threads participating in the multi-threaded scenario (in this case, producer and consumer) are blocked. If so, it increments the beat. Thus sometime later the beat will be incremented, from 0 to 1. Because consumer was waiting for beat 1, it will wake up (i.e., the waitForBeat(1) call will return) and execute the next line of code in its block, buf.take should be (42). This will succeed, because the producer thread had previously (during beat 0) put 42 into the queue. This act will also make producer runnable again, because it was blocked on the second put, which was waiting for another thread to read that 42.

   Now both threads are unblocked and able to execute their next statement. The order is non-deterministic, and can even be simultaneous if running on multiple cores. If the consumer thread happens to execute buf.take should be (17) first, it will block (buf.take will not return), because the queue is at that point empty. At some point later, the producer thread will execute buf put 17, which will unblock the consumer thread. Again both threads will be runnable and the order non-deterministic and possibly simulataneous. The producer thread may charge ahead and run its next statement, beat should be (1). This will succeed because the beat is indeed 1 at this point. As this is the last statement in the producer's block, the producer thread will exit normally (it won't throw an exception). At some point later the consumer thread will be allowed to complete its last statement, the buf.take call will return 17. The consumer thread will execute 17 should be (17). This will succeed and as this was the last statement in its block, the consumer will return normally.

   If either the producer or consumer thread had completed abruptbly with an exception, the conduct method (which was called by whenFinished) would have completed abruptly with an exception to indicate the test failed. However, since both threads returned normally, conduct will return. Because conduct doesn't throw an exception, whenFinished will execute the block of code passed as a by-name parameter to it: buf should be ('empty). This will succeed, because the queue is indeed empty at this point. The whenFinished method will then return, and because the whenFinished call was the last statement in the test and it didn't throw an exception, the test completes successfully.

   This test tests ArrayBlockingQueue, to make sure it works as expected. If there were a bug in ArrayBlockingQueue such as a put called on a full queue didn't block, but instead overwrote the previous value, this test would detect it. However, if there were a bug in ArrayBlockingQueue such that a call to take called on an empty queue never blocked and always returned 0, this test might not detect it. The reason is that whether the consumer thread will ever call take on an empty queue during this test is non-deterministic. It depends on how the threads get scheduled during beat 1. What is deterministic in this test, because the consumer thread blocks during beat 0, is that the producer thread will definitely attempt to write to a full queue. To make sure the other scenario is tested, you'd need a different test:

   test("calling take on an empty queue blocks the consumer thread") {
   
     val conductor = new Conductor
     import conductor._
   
     val buf = new ArrayBlockingQueue[Int](1)
   
     thread("producer") {
       waitForBeat(1)
       buf put 42
       buf put 17
     }
   
     thread("consumer") {
       buf.take should be (42)
       buf.take should be (17)
       beat should be (1)
     }
   
     whenFinished {
       buf should be ('empty)
     }
   }
   

   In this test, the producer thread will block, waiting for beat 1. The consumer thread will invoke buf.take as its first act. This will block, because the queue is empty. Because both threads are blocked, the Conductor will at some point later increment the beat to 1. This will awaken the producer thread. It will return from its waitForBeat(1) call and execute buf put 42. This will unblock the consumer thread, which will take the 42, and so on.

   The problem that Conductor is designed to address is the difficulty, caused by the non-deterministic nature of thread scheduling, of testing classes, traits, and libraries that are intended to be used by multiple threads. If you just create a test in which one thread reads from an ArrayBlockingQueue and another writes to it, you can't be sure that you have tested all possible interleavings of threads, no matter how many times you run the test. The purpose of Conductor is to enable you to write tests with deterministic interleavings of threads. If you write one test for each possible interleaving of threads, then you can be sure you have all the scenarios tested. The two tests shown here, for example, ensure that both the scenario in which a producer thread tries to write to a full queue and the scenario in which a consumer thread tries to take from an empty queue are tested.

   Class Conductor was inspired by the MultithreadedTC project, created by Bill Pugh and Nat Ayewah of the University of Maryland.

   Although useful, bear in mind that a Conductor's results are not guaranteed to be accurate 100% of the time. The reason is that it uses java.lang.Thread's getState method to decide when to advance the beat. This use goes against the advice given in the Javadoc documentation for getState, which says, "This method is designed for use in monitoring of the system state, not for synchronization." In short, sometimes the return value of getState occasionally may be inacurrate, which in turn means that sometimes a Conductor could decide to advance the beat too early. In practice, Conductor has proven to be very helpful when developing thread safe classes. It is also useful in for regression tests, but you may have to tolerate occasional false negatives.

6. trait Eventually extends PatienceConfiguration

   

   Trait that provides the eventually construct, which periodically retries executing a passed by-name parameter, until it either succeeds or the configured timeout has been surpassed.

   Trait that provides the eventually construct, which periodically retries executing a passed by-name parameter, until it either succeeds or the configured timeout has been surpassed.

   The by-name parameter "succeeds" if it returns a result. It "fails" if it throws any exception that would normally cause a test to fail. (These are any exceptions except TestPendingException and Errors listed in the Treatment of java.lang.Errors section of the documentation of trait Suite.)

   For example, the following invocation of eventually would succeed (not throw an exception):

   val xs = 1 to 125
   val it = xs.iterator
   eventually { it.next should be (3) }
   

   However, because the default timeout is 150 milliseconds, the following invocation of eventually would ultimately produce a TestFailedDueToTimeoutException:

   val xs = 1 to 125
   val it = xs.iterator
   eventually { Thread.sleep(50); it.next should be (110) }
   

   Assuming the default configuration parameters, a timeout of 150 milliseconds and an interval of 15 milliseconds, were passed implicitly to eventually, the detail message of the thrown TestFailedDueToTimeoutException would look like:

   The code passed to eventually never returned normally. Attempted 2 times over 166.682 milliseconds. Last failure message: 2 was not equal to 110.

   The cause of the thrown TestFailedDueToTimeoutException will be the exception most recently thrown by the block of code passed to eventually. (In the previous example, the cause would be the TestFailedException with the detail message 2 was not equal to 100.)

   Configuration of eventually

   The eventually methods of this trait can be flexibly configured. The two configuration parameters for eventually along with their default values and meanings are described in the following table:

   Configuration Parameter	Default Value	Meaning
   timeout	scaled(150 milliseconds)	the maximum amount of time to allow unsuccessful attempts before giving up and throwing TestFailedDueToTimeoutException
   interval	scaled(15 milliseconds)	the amount of time to sleep between each attempt

   The default values of both timeout and interval are passed to the scaled method, inherited from ScaledTimeSpans, so that the defaults can be scaled up or down together with other scaled time spans. See the documentation for trait ScaledTimeSpans for more information.

   The eventually methods of trait Eventually each take a PatienceConfig object as an implicit parameter. This object provides values for the two configuration parameters. (These configuration parameters are called "patience" because they determine how patient tests will be with asynchronous operations: how long they will tolerate failures before giving up and how long they will wait before checking again after a failure.) Trait Eventually provides an implicit val named patienceConfig with each configuration parameter set to its default value. If you want to set one or more configuration parameters to a different value for all invocations of eventually in a suite you can override this val (or hide it, for example, if you are importing the members of the Eventually companion object rather than mixing in the trait). For example, if you always want the default timeout to be 2 seconds and the default interval to be 5 milliseconds, you can override patienceConfig, like this:

   implicit override val patienceConfig =
     PatienceConfig(timeout = scaled(Span(2, Seconds)), interval = scaled(Span(5, Millis)))
   

   Or, hide it by declaring a variable of the same name in whatever scope you want the changed values to be in effect:

   implicit val patienceConfig =
     PatienceConfig(timeout = scaled(Span(2, Seconds)), interval = scaled(Span(5, Millis)))
   

   Passing your new default values to scaled is optional, but a good idea because it allows them to be easily scaled if run on a slower or faster system.

   In addition to taking a PatienceConfig object as an implicit parameter, the eventually methods of trait Eventually include overloaded forms that take one or two PatienceConfigParam objects that you can use to override the values provided by the implicit PatienceConfig for a single eventually invocation. For example, if you want to set timeout to 5 seconds for just one particular eventually invocation, you can do so like this:

   eventually (timeout(Span(5, Seconds))) { Thread.sleep(10); it.next should be (110) }
   

   This invocation of eventually will use 5 seconds for the timeout and whatever value is specified by the implicitly passed PatienceConfig object for the interval configuration parameter. If you want to set both configuration parameters in this way, just list them separated by commas:

   eventually (timeout(Span(5, Seconds)), interval(Span(5, Millis))) { it.next should be (110) }
   

   You can also import or mix in the members of SpanSugar if you want a more concise DSL for expressing time spans:

   eventually (timeout(5 seconds), interval(5 millis)) { it.next should be (110) }
   

   Note that ScalaTest will not scale any time span that is not explicitly passed to scaled to make the meaning of the code as obvious as possible. Thus if you ask for "timeout(5 seconds)" you will get exactly that: a timeout of five seconds. If you want such explicitly given values to be scaled, you must pass them to scale explicitly like this:

   eventually (timeout(scaled(5 seconds))) { it.next should be (110) }
   

   The previous code says more clearly that the timeout will be five seconds, unless scaled higher or lower by the scaled method.

   Simple backoff algorithm

   The eventually methods employ a very simple backoff algorithm to try and maximize the speed of tests. If an asynchronous operation completes quickly, a smaller interval will yield a faster test. But if an asynchronous operation takes a while, a small interval will keep the CPU busy repeatedly checking and rechecking a not-ready operation, to some extent taking CPU cycles away from other processes that could proceed. To strike the right balance between these design tradeoffs, the eventually methods will check more frequently during the initial interval.

   Rather than sleeping an entire interval if the initial attempt fails, eventually will only sleep 1/10 of the configured interval. It will continue sleeping only 1/10 of the configured interval until the configured interval has passed, after which it sleeps the configured interval between attempts. Here's an example in which the timeout is set equal to the interval:

   val xs = 1 to 125
   val it = xs.iterator
   eventually(timeout(100 milliseconds), interval(100 milliseconds)) { it.next should be (110) }
   

   Even though this call to eventually will time out after only one interval, approximately, the error message will likely report that more than one (and less than ten) attempts were made:

   The code passed to eventually never returned normally. Attempted 6 times over 100.485 milliseconds. Last failure message: 6 was not equal to 110.

   Note that if the initial attempt takes longer than the configured interval to complete, eventually will never sleep for a 1/10 interval. You can observe this behavior in the second example above in which the first statement in the block of code passed to eventually was Thread.sleep(50).

   Usage note: Eventually intended primarily for integration testing

   Although the default timeouts of trait Eventually are tuned for unit testing, the use of Eventually in unit tests is a choice you should question. Usually during unit testing you'll want to mock out subsystems that would require Eventually, such as network services with varying and unpredictable response times. This will allow your unit tests to run as fast as possible while still testing the focused bits of behavior they are designed to test.

   Nevertheless, because sometimes it will make sense to use Eventually in unit tests (and because it is destined to happen anyway even when it isn't the best choice), Eventually by default uses timeouts tuned for unit tests: Calls to eventually are more likely to succeed on fast development machines, and if a call does time out, it will do so quickly so the unit tests can move on.

   When you are using Eventually for integration testing, therefore, the default timeout and interval may be too small. A good way to override them is by mixing in trait IntegrationPatience or a similar trait of your own making. Here's an example:

   class ExampleSpec extends FeatureSpec with Eventually with IntegrationPatience {
     // Your integration tests here...
   }
   

   Trait IntegrationPatience increases the default timeout from 150 milliseconds to 15 seconds, the default interval from 15 milliseconds to 150 milliseconds. If need be, you can do fine tuning of the timeout and interval by specifying a time span scale factor when you run your tests.

7. trait Futures extends PatienceConfiguration

   

   Trait that facilitates testing with futures.

   Trait that facilitates testing with futures.

   This trait defines a FutureConcept trait that can be used to implicitly wrap different kinds of futures, thereby providing a uniform testing API for futures. The three ways this trait enables you to test futures are:

   1. Invoking isReadyWithin, to assert that a future is ready within a a specified time period. Here's an example:

   assert(result.isReadyWithin(100 millis))
   

   2. Invoking futureValue, to obtain a futures result within a specified or implicit time period, like this:

   assert(result.futureValue === 7)
   

   3. Passing the future to whenReady, and performing assertions on the result value passed to the given function, as in:

   whenReady(result) { s =>
     s should be ("hello")
   }
   

   The whenReady construct periodically inspects the passed future, until it is either ready or the configured timeout has been surpassed. If the future becomes ready before the timeout, whenReady passes the future's value to the specified function.

   To make whenReady more broadly applicable, the type of future it accepts is a FutureConcept[T], where T is the type of value promised by the future. Passing a future to whenReady requires an implicit conversion from the type of future you wish to pass (the modeled type) to FutureConcept[T]. Subtrait JavaFutures provides an implicit conversion from java.util.concurrent.Future[T] to FutureConcept[T].

   For example, the following invocation of whenReady would succeed (not throw an exception):

   import org.scalatest._
   import matchers.ShouldMatchers._
   import concurrent.Futures._
   import java.util.concurrent._
   
   val exec = Executors.newSingleThreadExecutor
   val task = new Callable[String] { def call() = { Thread.sleep(50); "hi" } }
   whenReady(exec.submit(task)) { s =>
     s should be ("hi")
   }
   

   However, because the default timeout is 150 milliseconds, the following invocation of whenReady would ultimately produce a TestFailedException:

   val task = new Callable[String] { def call() = { Thread.sleep(500); "hi" } }
   whenReady(exec.submit(task)) { s =>
     s should be ("hi")
   }
   

   Assuming the default configuration parameters, a timeout of 150 milliseconds and an interval of 15 milliseconds, were passed implicitly to whenReady, the detail message of the thrown TestFailedException would look like:

   The future passed to whenReady was never ready, so whenReady timed out. Queried 95 times, sleeping 10 milliseconds between each query.

   Configuration of whenReady

   The whenReady methods of this trait can be flexibly configured. The two configuration parameters for whenReady along with their default values and meanings are described in the following table:

   Configuration Parameter	Default Value	Meaning
   timeout	scaled(150 milliseconds)	the maximum amount of time to allow unsuccessful queries before giving up and throwing TestFailedException
   interval	scaled(15 milliseconds)	the amount of time to sleep between each query

   The default values of both timeout and interval are passed to the scaled method, inherited from ScaledTimeSpans, so that the defaults can be scaled up or down together with other scaled time spans. See the documentation for trait ScaledTimeSpans for more information.

   The whenReady methods of trait Futures each take a PatienceConfig object as an implicit parameter. This object provides values for the two configuration parameters. Trait Futures provides an implicit val named defaultPatience with each configuration parameter set to its default value. If you want to set one or more configuration parameters to a different value for all invocations of whenReady in a suite you can override this val (or hide it, for example, if you are importing the members of the Futures companion object rather than mixing in the trait). For example, if you always want the default timeout to be 2 seconds and the default interval to be 5 milliseconds, you can override defaultPatience, like this:

   implicit override val defaultPatience =
     PatienceConfig(timeout = Span(2, Seconds), interval = Span(5, Millis))
   

   Or, hide it by declaring a variable of the same name in whatever scope you want the changed values to be in effect:

   implicit val defaultPatience =
     PatienceConfig(timeout =  Span(2, Seconds), interval = Span(5, Millis))
   

   In addition to taking a PatienceConfig object as an implicit parameter, the whenReady methods of trait Futures include overloaded forms that take one or two PatienceConfigParam objects that you can use to override the values provided by the implicit PatienceConfig for a single whenReady invocation. For example, if you want to set timeout to 6 seconds for just one particular whenReady invocation, you can do so like this:

   whenReady (exec.submit(task), timeout(Span(6, Seconds))) { s =>
     s should be ("hi")
   }
   

   This invocation of eventually will use 6000 for timeout and whatever value is specified by the implicitly passed PatienceConfig object for the interval configuration parameter. If you want to set both configuration parameters in this way, just list them separated by commas:

   whenReady (exec.submit(task), timeout(Span(6, Seconds)), interval(Span(500, Millis))) { s =>
     s should be ("hi")
   }
   

   You can also import or mix in the members of SpanSugar if you want a more concise DSL for expressing time spans:

   whenReady (exec.submit(task), timeout(6 seconds), interval(500 millis)) { s =>
     s should be ("hi")
   }
   

   Note: The whenReady construct was in part inspired by the whenDelivered matcher of the BlueEyes project, a lightweight, asynchronous web framework for Scala.

8. trait IntegrationPatience extends AbstractPatienceConfiguration

   

   Stackable modification trait for PatienceConfiguration that provides default timeout and interval values appropriate for integration testing.

   Stackable modification trait for PatienceConfiguration that provides default timeout and interval values appropriate for integration testing.

   The default values for the parameters are:

   Configuration Parameter	Default Value
   timeout	scaled(15 seconds)
   interval	scaled(150 milliseconds)

   The default values of both timeout and interval are passed to the scaled method, inherited from ScaledTimeSpans, so that the defaults can be scaled up or down together with other scaled time spans. See the documentation for trait ScaledTimeSpans for more information.

   Mix this trait into any class that uses PatienceConfiguration (such as classes that mix in Eventually or AsyncAssertions) to get timeouts tuned towards integration testing, like this:

   class ExampleSpec extends FeatureSpec with Eventually with IntegrationPatience {
     // ...
   }
   

9. trait Interruptor extends (Thread) ⇒ Unit

   

   Strategy for interrupting an operation after a timeout expires.

   Strategy for interrupting an operation after a timeout expires.

   An instance of this trait is used for configuration when using traits Timeouts and TimeLimitedTests.

10. trait JavaFutures extends Futures

    

    Provides an implicit conversion from java.util.concurrent.Future[T] to FutureConcept[T].

    Provides an implicit conversion from java.util.concurrent.Future[T] to FutureConcept[T].

    This trait enables you to invoke the methods defined on FutureConcept on a Java Future, as well as to pass a Java future to the whenReady methods of supertrait Futures. See the documentation for supertrait Futures for the details on the syntax this trait provides for testing with Java futures.

11. trait PatienceConfiguration extends AbstractPatienceConfiguration

    

    Trait providing methods and classes used to configure timeouts and, where relevant, the interval between retries.

    Trait providing methods and classes used to configure timeouts and, where relevant, the interval between retries.

    This trait is called PatienceConfiguration because it allows configuration of two values related to patience: The timeout specifies how much time asynchronous operations will be given to succeed before giving up. The interval specifies how much time to wait between checks to determine success when polling.

    The default values for timeout and interval provided by trait PatienceConfiguration are tuned for unit testing, where running tests as fast as possible is a high priority and subsystems requiring asynchronous operations are therefore often replaced by mocks. This table shows the default values:

    Configuration Parameter	Default Value
    timeout	scaled(150 milliseconds)
    interval	scaled(15 milliseconds)

    Values more appropriate to integration testing, where asynchronous operations tend to take longer because the tests are run against the actual subsytems (not mocks), can be obtained by mixing in trait IntegrationPatience.

    The default values of both timeout and interval are passed to the scaled method, inherited from ScaledTimeSpans, so that the defaults can be scaled up or down together with other scaled time spans. See the documentation for trait ScaledTimeSpans for more information.

    Timeouts are used by the eventually methods of trait Eventually and the await method of class Waiter, a member of trait AsyncAssertions. Intervals are used by the eventually methods.

12. trait ScalaFutures extends Futures

    

    Provides an implicit conversion from scala.concurrent.Future[T] to FutureConcept[T].

    Provides an implicit conversion from scala.concurrent.Future[T] to FutureConcept[T].

    This trait enables you to invoke the methods defined on FutureConcept on a Scala Future, as well as to pass a Scala future to the whenReady methods of supertrait Futures. See the documentation for supertrait Futures for the details on the syntax this trait provides for testing with Scala futures.

13. trait ScaledTimeSpans extends AnyRef

    

    Trait providing a scaled method that can be used to scale time Spans used during the testing of asynchronous operations.

    Trait providing a scaled method that can be used to scale time Spans used during the testing of asynchronous operations.

    The scaled method allows tests of asynchronous operations to be tuned according to need. For example, Spans can be scaled larger when running tests on slower continuous integration servers or smaller when running on faster development machines.

    The Double factor by which to scale the Spans passed to scaled is obtained from the spanScaleFactor method, also declared in this trait. By default this method returns 1.0, but can be configured to return a different value by passing a -F argument to Runner (or an equivalent mechanism in an ant, sbt, or Maven build file).

    The default timeouts and intervals defined for traits Eventually and AsyncAssertions invoke scaled, so those defaults will be scaled automatically. Other than such defaults, however, to get a Span to scale you'll need to explicitly pass it to scaled. For example, here's how you would scale a Span you supply to the failAfter method from trait Timeouts:

    failAfter(scaled(150 millis)) {
      // ...
    }
    

    The reason Spans are not scaled automatically in the general case is to make code obvious. If a reader sees failAfter(1 second), it will mean exactly that: fail after one second. And if a Span will be scaled, the reader will clearly see that as well: failAfter(scaled(1 second)).

    Overriding spanScaleFactor

    You can override the spanScaleFactor method to configure the factor by a different means. For example, to configure the factor from Akka TestKit's test time factor you might create a trait like this:

    import org.scalatest.concurrent.ScaledTimeSpans
    import akka.actor.ActorSystem
    import akka.testkit.TestKitExtension
    
    trait AkkaSpanScaleFactor extends ScaledTimeSpans {
      override def spanScaleFactor: Double =
          TestKitExtension.get(ActorSystem()).TestTimeFactor
    }
    

    This trait overrides spanScaleFactor so that it takes its scale factor from Akka's application.conf file. You could then scale Spans tenfold in Akka's configuration file like this:

    akka {
      test {
        timefactor = 10.0
      }
    }
    

    Armed with this trait and configuration file, you can simply mix trait AkkaSpanScaleFactor into any test class whose Spans you want to scale, like this:

    class MySpec extends FunSpec with Eventually with AkkaSpanScaleFactor {
      // ..
    }
    

14. class SelectorInterruptor extends Interruptor

    

    Strategy for interrupting an operation in which wakeup is called on the java.nio.channels.Selector passed to the constructor.

    Strategy for interrupting an operation in which wakeup is called on the java.nio.channels.Selector passed to the constructor.

    This class can be used for configuration when using traits Timeouts and TimeLimitedTests.

15. class SocketInterruptor extends Interruptor

    

    Strategy for interrupting an operation in which close is called on the java.net.Socket passed to the constructor.

    Strategy for interrupting an operation in which close is called on the java.net.Socket passed to the constructor.

    This class can be used for configuration when using traits Timeouts and TimeLimitedTests.

16. trait TimeLimitedTests extends SuiteMixin

    

    Trait that when mixed into a suite class establishes a time limit for its tests.

    Trait that when mixed into a suite class establishes a time limit for its tests.

    This trait overrides withFixture, wrapping a super.withFixture(test) call in a failAfter invocation, specifying a timeout obtained by invoking timeLimit and an Interruptor by invoking defaultTestInterruptor:

    failAfter(timeLimit) {
      super.withFixture(test)
    } (defaultTestInterruptor)
    

    Note that the failAfter method executes the body of the by-name passed to it using the same thread that invoked failAfter. This means that the same thread will run the withFixture method as well as each test, so no extra synchronization is required. A second thread is used to run a timer, and if the timeout expires, that second thread will attempt to interrupt the main test thread via the defaultTestInterruptor.

    The timeLimit field is abstract in this trait. Thus you must specify a time limit when you use it. For example, the following code specifies that each test must complete within 200 milliseconds:

    import org.scalatest.FunSpec
    import org.scalatest.concurrent.TimeLimitedTests
    import org.scalatest.time.SpanSugar._
    
    class ExampleSpec extends FunSpec with TimeLimitedTests {
    
      // Note: You may need to either write 200.millis or (200 millis), or
      // place a semicolon or blank line after plain old 200 millis, to
      // avoid the semicolon inference problems of postfix operator notation.
      val timeLimit = 200 millis
    
      describe("A time-limited test") {
        it("should succeed if it completes within the time limit") {
          Thread.sleep(100)
        }
        it("should fail if it is taking too darn long") {
          Thread.sleep(300)
        }
      }
    }
    

    If you run the above ExampleSpec, the second test will fail with the error message:

    The test did not complete within the specified 200 millisecond time limit.

    The failAfter method uses an Interruptor to attempt to interrupt the main test thread if the timeout expires. The default Interruptor returned by the defaultTestInterruptor method is a ThreadInterruptor, which calls interrupt on the main test thread. If you wish to change this interruption strategy, override defaultTestInterruptor to return a different Interruptor. For example, here's how you'd change the default to DoNotInterrupt, a very patient interruption strategy that does nothing to interrupt the main test thread:

    import org.scalatest.FunSpec
    import org.scalatest.concurrent.TimeLimitedTests
    import org.scalatest.time.SpanSugar._
    
    class ExampleSpec extends FunSpec with TimeLimitedTests {
    
      val timeLimit = 200 millis
    
      override val defaultTestInterruptor = DoNotInterrupt
    
      describe("A time-limited test") {
        it("should succeed if it completes within the time limit") {
          Thread.sleep(100)
        }
        it("should fail if it is taking too darn long") {
          Thread.sleep(300)
        }
      }
    }
    

    Like the previous incarnation of ExampleSuite, the second test will fail with an error message that indicates a timeout expired. But whereas in the previous case, the Thread.sleep would be interrupted after 200 milliseconds, in this case it is never interrupted. In the previous case, the failed test requires a little over 200 milliseconds to run. In this case, because the sleep(300) is never interrupted, the failed test requires a little over 300 milliseconds to run.

17. trait Timeouts extends AnyRef

    

    Trait that provides a failAfter and cancelAfter construct, which allows you to specify a time limit for an operation passed as a by-name parameter, as well as a way to interrupt it if the operation exceeds its time limit.

    Trait that provides a failAfter and cancelAfter construct, which allows you to specify a time limit for an operation passed as a by-name parameter, as well as a way to interrupt it if the operation exceeds its time limit.

    The time limit is passed as the first parameter, as a Span. The operation is passed as the second parameter. And an Interruptor, a strategy for interrupting the operation, is passed as an implicit third parameter. Here's a simple example of its use:

    failAfter(Span(100, Millis)) {
      Thread.sleep(200)
    }
    

    The above code, after 100 milliseconds, will produce a TestFailedDueToTimeoutException with a message that indicates a timeout expired:

    The code passed to failAfter did not complete within 100 milliseconds.

    If you use cancelAfter in place of failAfter, a TestCanceledException with a message that indicates a timeout expired:

    The code passed to cancelAfter did not complete within 100 milliseconds.

    If you prefer you can mix in or import the members of SpanSugar and place a units value after the integer timeout. Here are some examples:

    import org.scalatest.time.SpanSugar._
    
    failAfter(100 millis) {
      Thread.sleep(200)
    }
    
    failAfter(1 second) {
      Thread.sleep(2000)
    }
    

    The code passed via the by-name parameter to failAfter or cancelAfter will be executed by the thread that invoked failAfter or cancelAfter, so that no synchronization is necessary to access variables declared outside the by-name.

    var result = -1 // No need to make this volatile
    failAfter(100 millis) {
      result = accessNetService()
    }
    result should be (99)
    

    The failAfter or cancelAfter method will create a timer that runs on a different thread than the thread that invoked failAfter or cancelAfter, so that it can detect when the timeout has expired and attempt to interrupt the main thread. Because different operations can require different interruption strategies, the failAfter or cancelAfter method accepts an implicit third parameter of type Interruptor that is responsible for interrupting the main thread.

    Configuring failAfter or cancelAfter with an Interruptor

    This trait declares an implicit val named defaultInterruptor, initialized with a ThreadInterruptor, which attempts to interrupt the main thread by invoking Thread.interrupt. If you wish to use a different strategy, you can override this val (or hide it, for example if you imported the members of Timeouts rather than mixing it in). Here's an example in which the default interruption method is changed to DoNotInterrupt, which does not attempt to interrupt the main thread in any way:

    override val defaultInterruptor = DoNotInterrupt
    failAfter(100 millis) {
      Thread.sleep(500)
    }
    

    As with the default Interruptor, the above code will eventually produce a TestFailedDueToTimeoutException with a message that indicates a timeout expired. However, instead of throwing the exception after approximately 100 milliseconds, it will throw it after approximately 500 milliseconds.

    This illustrates an important feature of failAfter and cancelAfter: it will throw a TestFailedDueToTimeoutException (or TestCanceledException in case of cancelAfter) if the code passed as the by-name parameter takes longer than the specified timeout to execute, even if it is allowed to run to completion beyond the specified timeout and returns normally.

    ScalaTest provides the following Interruptor implementations:

    Interruptor implementation	Usage
    ThreadInterruptor	The default interruptor, invokes interrupt on the main test thread. This will set the interrupted status for the main test thread and, if the main thread is blocked, will in some cases cause the main thread to complete abruptly with an InterruptedException.
    DoNotInterrupt	Does not attempt to interrupt the main test thread in any way
    SelectorInterruptor	Invokes wakeup on the passed java.nio.channels.Selector, which will cause the main thread, if blocked in Selector.select, to complete abruptly with a ClosedSelectorException.
    SocketInterruptor	Invokes close on the java.io.Socket, which will cause the main thread, if blocked in a read or write of an java.io.InputStream or java.io.OutputStream that uses the Socket, to complete abruptly with a SocketException.

    You may wish to create your own Interruptor in some situations. For example, if your operation is performing a loop and can check a volatile flag each pass through the loop. You could in that case write an Interruptor that sets that flag so that the next time around, the loop would exit.