LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

       * UnsignedBytesBenchmark. TODO(cpovirk): benchmark this
       */
      /*
       * A CopyOnWriteArraySet<WeakReference> is faster than a newSetFromMap of a MapMaker map with
       * weakKeys() and concurrencyLevel(1), even up to at least 12 cached exception types.
       */
      private static final Set<WeakReference<Class<? extends Exception>>> validClasses =
          new CopyOnWriteArraySet<>();

      @Override

       * timeoutFuture can be set (and delegate cancelled) only once. (And "set only once" is
       * important for other reasons: run() can still be invoked concurrently in different threads,
       * even with the above null checks.)
       */
      timeoutFutureRef = null;
      if (delegate.isDone()) {
        timeoutFuture.setFuture(delegate);
      } else {
        try {

   *       uppercase, hexadecimal representation of the byte value.
   * </ul>
   *
   * <p>This escaper is suitable for escaping parameter names and values even when <a
   * href="https://www.w3.org/TR/html401/appendix/notes.html#h-B.2.2">using the non-standard
   * semicolon</a>, rather than the ampersand, as a parameter delimiter. Nevertheless, we recommend

 * bound. However, it would also let them create `new TypeParameter<@Nullable T>() {}`, which
 * wouldn't behave as users might expect. Additionally, it's not clear how the TypeToken API could
 * support even a "normal" `TypeParameter<T>` when `<T>` has a nullable bound. (See the discussion
 * on TypeToken.where.) So, in the interest of failing fast and encouraging the user to switch to a

    tests, either. This causes it to fail to find AtomicLongMapTest.

    Our workaround is to tell GWT that util.concurrent and all other packages
    have prod supersource, even if they have none. GWT is happy to ignore us
    when we specify a nonexistent path.

    (I hope that this workaround does not cause its own problems in the future.)
-->
<super-source path="super"/>

  }

  public void testBroken() {
    Iterator<Object> broken = new BrokenAbstractSequentialIterator();
    assertTrue(broken.hasNext());
    // We can't retrieve even the known first element:
    assertThrows(SomeUncheckedException.class, broken::next);
    assertThrows(SomeUncheckedException.class, broken::next);
  }

 * bound. However, it would also let them create `new TypeParameter<@Nullable T>() {}`, which
 * wouldn't behave as users might expect. Additionally, it's not clear how the TypeToken API could
 * support even a "normal" `TypeParameter<T>` when `<T>` has a nullable bound. (See the discussion
 * on TypeToken.where.) So, in the interest of failing fast and encouraging the user to switch to a

   * rather than just the {@link RuntimeException} or {@link Error} that should be possible. (And in
   * fact the static type of {@link Throwable} is occasionally justified even for a method with no
   * {@code throws} clause: Some such methods can in fact throw a checked exception (e.g., by
   * calling code written in Kotlin).) Typically, we want to let a {@link Throwable} from such a

   * rather than just the {@link RuntimeException} or {@link Error} that should be possible. (And in
   * fact the static type of {@link Throwable} is occasionally justified even for a method with no
   * {@code throws} clause: Some such methods can in fact throw a checked exception (e.g., by
   * calling code written in Kotlin).) Typically, we want to let a {@link Throwable} from such a