Changed the test case to produce a reliable predictable result (#320)

modules/hcc-model/src/test/java/net/sf/dz3r/device/actuator/economizer/v2/PidEconomizerTest.java

@@ -1,5 +1,6 @@
 package net.sf.dz3r.device.actuator.economizer.v2;
 
+import net.sf.dz3r.device.DeviceState;
 import net.sf.dz3r.device.actuator.NullCqrsSwitch;
 import net.sf.dz3r.device.actuator.SwitchableHvacDevice;
 import net.sf.dz3r.device.actuator.economizer.EconomizerConfig;
@@ -8,12 +9,13 @@
 import net.sf.dz3r.signal.Signal;
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
-import org.junit.jupiter.api.RepeatedTest;
+import org.junit.jupiter.api.Test;
 import reactor.core.publisher.Sinks;
 import reactor.core.scheduler.Schedulers;
 
 import java.time.Clock;
 import java.time.Duration;
+import java.util.concurrent.LinkedBlockingQueue;
 import java.util.concurrent.TimeoutException;
 import java.util.concurrent.atomic.AtomicLong;
 
@@ -28,15 +30,15 @@ class PidEconomizerTest {
      *
      * See <a href="https://github.com/home-climate-control/dz/issues/320">#320</a>.
      */
-    @RepeatedTest(value = 10, failureThreshold = 1)
+    @Test
     void ambientSensorLoss() throws InterruptedException {
 
-        var s = new NullCqrsSwitch("test");
+        var actuator = new NullCqrsSwitch("test");
         var hvac = new SwitchableHvacDevice(
                 Clock.systemUTC(),
                 "test",
                 HvacMode.COOLING,
-                s,
+                actuator,
                 false,
                 null
         );
@@ -76,23 +78,24 @@ void ambientSensorLoss() throws InterruptedException {
                 .publishOn(Schedulers.boundedElastic())
                 .subscribe(step -> logger.info("step: {}", step));
 
+        var deviceState = new LinkedBlockingQueue<Signal<DeviceState<Boolean>, String>>();
+
+        actuator.getFlux().subscribe(deviceState::add);
+
         // Setup complete, let's push data now
 
-        // Initially, the economizer is on
+        // After receiving just the indoor signal (but not ambient), the economizer is off
         indoorSink.tryEmitNext(new Signal<>(start, 25.0, "indoor"));
-        ambientSink.tryEmitNext(new Signal<>(start.plus(Duration.ofMillis(offset.addAndGet(timeStep))), 20.0));
+        assertThat(deviceState.take().getValue().requested).isFalse();
 
-        Thread.sleep(10); // NOSONAR Risks have been assessed and accepted
-        assertThat(s.getState().requested).isTrue();
+        // When both indoor and ambient signals are available, it is now on
+        ambientSink.tryEmitNext(new Signal<>(start.plus(Duration.ofMillis(offset.addAndGet(timeStep))), 20.0));
+        assertThat(deviceState.take().getValue().requested).isTrue();
 
         // It should turn off now
         ambientSink.tryEmitNext(new Signal<>(start.plus(Duration.ofMillis(offset.addAndGet(timeStep))), null, null,Signal.Status.FAILURE_TOTAL, new TimeoutException("oops")));
 
-        Thread.sleep(10); // NOSONAR Risks have been assessed and accepted
-
-        // As of rev. 139c3ae65438f4a98c90e74dbffe24073e178d20:
-        // If at any time the I component of stage 1 controller signal is more than 0, this makes its output less than the renderer threshold because the value of -1 is used for signaling,
-        // and the economizer gets stuck on.
-        assertThat(s.getState().requested).isFalse();
+        // ...and it does.
+        assertThat(deviceState.take().getValue().requested).isFalse();
     }
 }