@NullUnmarked
public class NullCacheTest extends TestCase {
  QueuingRemovalListener<Object, Object> listener;

  @Override
  protected void setUp() {
    listener = queuingRemovalListener();
  }

  public void testGet() {
    Object computed = new Object();
    LoadingCache<Object, Object> cache =
        CacheBuilder.newBuilder()
            .maximumSize(0)
            .removalListener(listener)

        assertThat(map.containsKey("one"), is(true));
        assertThat(map.containsKey("onex"), is(not(true)));
    }

    /**
     * @throws Exception
     */
    @Test
    public void testGet() throws Exception {
        assertThat(map.get("ONE"), is("1"));
        assertThat(map.get("One"), is("1"));
        assertThat(map.get("hoge"), is(nullValue()));
    }

    /**
     * @throws Exception
     */

Ayrıca nelerin değiştiğini, validasyonların artık nasıl daha doğru ve katı olduğunu, olası dikkat edilmesi gereken noktaları (caveat) vb. de içerir.

Nelerin değiştiğini daha iyi anlamak için okuyabilirsiniz.

## Testler { #tests }

Uygulamanız için [testleriniz](../tutorial/testing.md) olduğundan ve bunları continuous integration (CI) üzerinde çalıştırdığınızdan emin olun.

import com.google.common.collect.testing.features.CollectionSize;
import com.google.common.collect.testing.features.MapFeature;
import java.util.Iterator;
import org.junit.Ignore;

/**
 * Tester for {@code Multimap.entries}.
 *
 * @author Louis Wasserman
 */
@GwtCompatible
@Ignore("test runners must not instantiate and run this directly, only via suites we build")

Insbesondere dass Uvicorn, Starlette und FastAPI zusammen verglichen werden (neben vielen anderen Tools).

Je einfacher das Problem, das durch das Tool gelöst wird, desto besser wird die Performanz sein. Und die meisten Benchmarks testen nicht die zusätzlichen Funktionen, die das Tool bietet.

Die Hierarchie ist wie folgt:

* **Uvicorn**: ein ASGI-Server
    * **Starlette**: (verwendet Uvicorn) ein Web-Mikroframework

import com.google.common.collect.testing.features.CollectionSize;
import com.google.common.collect.testing.features.MapFeature;
import org.junit.Ignore;

/**
 * Tester for {@code Multimap.toString()}.
 *
 * @author Louis Wasserman
 */
@GwtCompatible
@Ignore("test runners must not instantiate and run this directly, only via suites we build")

import com.google.common.collect.testing.features.CollectionSize;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
import java.util.Map.Entry;
import org.junit.Ignore;

/**
 * Tester for {@code Multimap.values}.
 *
 * @author Louis Wasserman
 */
@GwtCompatible
@Ignore("test runners must not instantiate and run this directly, only via suites we build")

      }
    }
  }

  public void testNulls() {
    NullPointerTester tester =
        new NullPointerTester()
            .setDefault(SleepingStopwatch.class, stopwatch)
            .setDefault(int.class, 1)
            .setDefault(double.class, 1.0d);
    tester.testStaticMethods(RateLimiter.class, Visibility.PACKAGE);
    tester.testInstanceMethods(RateLimiter.create(5.0, stopwatch), Visibility.PACKAGE);
  }

      }
    }
  }

  public void testNulls() {
    NullPointerTester tester =
        new NullPointerTester()
            .setDefault(SleepingStopwatch.class, stopwatch)
            .setDefault(int.class, 1)
            .setDefault(double.class, 1.0d);
    tester.testStaticMethods(RateLimiter.class, Visibility.PACKAGE);
    tester.testInstanceMethods(RateLimiter.create(5.0, stopwatch), Visibility.PACKAGE);
  }

 *       have, of returning a new, independent iterator
 * </ul>
 *
 * <p>Because of this situation, any public method accepting an iterable should invoke the {@code
 * iterator} method only once, and should be tested using this class. Exceptions to this rule should
 * be clearly documented.
 *
 * <p>Note that although your APIs should be liberal in what they accept, your methods which