(v1, v2) -> {
                  throw new IllegalArgumentException("Multiple values for key: " + v1 + ", " + v2);
                }),
        (accum, t) -> {
          /*
           * We assign these to variables before calling checkNotNull to work around a bug in our
           * nullness checker.
           */
          K key = keyFunction.apply(t);
          V newValue = valueFunction.apply(t);

    Ordering<Number> j = objects.compound(numbers.compound(objects)); // bad IDEA
    Ordering<Number> k = objects.compound(objects.compound(numbers));

    // You can also arbitrarily assign a more restricted type - not an intended
    // feature, exactly, but unavoidable (I think) and harmless
    Ordering<Integer> l = objects.compound(numbers);

    // This correctly doesn't work:

    Ordering<Number> j = objects.compound(numbers.compound(objects)); // bad IDEA
    Ordering<Number> k = objects.compound(objects.compound(numbers));

    // You can also arbitrarily assign a more restricted type - not an intended
    // feature, exactly, but unavoidable (I think) and harmless
    Ordering<Integer> l = objects.compound(numbers);

    // This correctly doesn't work:

      this.hashFunction = new ChecksumHashFunction(this, 32, toString);
    }
  }

  /**
   * Returns a hash function implementing FarmHash's Fingerprint64, an open-source algorithm.
   *
   * <p>This is designed for generating persistent fingerprints of strings. It isn't
   * cryptographically secure, but it produces a high-quality hash with fewer collisions than some
   * alternatives we've used in the past.
   *

   * or extract a variable, and put the suppression on that. However, a local variable typically
   * doesn't work: Because nullness analyses typically infer the nullness of local variables,
   * there's no way to assign a {@code @Nullable T} to a field {@code T foo;} and instruct the
   * analysis that that means "plain {@code T}" rather than the inferred type {@code @Nullable T}.

   * Returns the innermost cause of {@code throwable}. The first throwable in a chain provides
   * context from when the error or exception was initially detected. Example usage:
   *
   * <pre>
   * assertEquals("Unable to assign a customer id", Throwables.getRootCause(e).getMessage());
   * </pre>
   *
   * @throws IllegalArgumentException if there is a loop in the causal chain
   */
  public static Throwable getRootCause(Throwable throwable) {

   * Returns the innermost cause of {@code throwable}. The first throwable in a chain provides
   * context from when the error or exception was initially detected. Example usage:
   *
   * <pre>
   * assertEquals("Unable to assign a customer id", Throwables.getRootCause(e).getMessage());
   * </pre>
   *
   * @throws IllegalArgumentException if there is a loop in the causal chain
   */
  public static Throwable getRootCause(Throwable throwable) {

  // A note on memory visibility.
  // Many of the utilities in this class (transform, withFallback, withTimeout, asList, combine)
  // have two requirements that significantly complicate their design.
  // 1. Cancellation should propagate from the returned future to the input future(s).
  // 2. The returned futures shouldn't unnecessarily 'pin' their inputs after completion.
  //

   * or extract a variable, and put the suppression on that. However, a local variable typically
   * doesn't work: Because nullness analyses typically infer the nullness of local variables,
   * there's no way to assign a {@code @Nullable T} to a field {@code T foo;} and instruct the
   * analysis that that means "plain {@code T}" rather than the inferred type {@code @Nullable T}.

  }

  /*
   * To avoid the need to call requireNonNull so much, we could consider more clever approaches,
   * such as:
   *
   * - Make checkOpen return the non-null `seq`. Then callers can assign that to a local variable or
   *   even back to `this.seq`. However, that may suggest that we're defending against concurrent
   *   mutation, which is not an actual risk because we use `synchronized`.