@Override
  public int hashCode() {
    return Sets.hashCodeImpl(this);
  }

  // This declaration is needed to make Set.iterator() and
  // ImmutableCollection.iterator() appear consistent to javac's type inference.
  @Override
  public abstract UnmodifiableIterator<E> iterator();

  abstract static class CachingAsList<E> extends ImmutableSet<E> {
    @LazyInit private transient ImmutableList<E> asList;

    @Override

        return set1.contains(element) ^ set2.contains(element);
      }

      @Override
      int minSize() {
        int difference = minSize(set1) - maxSize(set2);
        return difference >= 0 ? difference : max(minSize(set2) - maxSize(set1), 0);
      }

      @Override
      int maxSize() {
        return saturatedAdd(maxSize(set1), maxSize(set2));
      }

        return set1.contains(element) ^ set2.contains(element);
      }

      @Override
      int minSize() {
        int difference = minSize(set1) - maxSize(set2);
        return difference >= 0 ? difference : max(minSize(set2) - maxSize(set1), 0);
      }

      @Override
      int maxSize() {
        return saturatedAdd(maxSize(set1), maxSize(set2));
      }

              getLowerEndpoint(),
              getLowerBoundType());
      result.reverse = this;
      return this.reverse = result;
    }
    return result;
  }

  // This method helps J2KT's type inference.
  private static <T extends @Nullable Object> Comparator<T> reverseComparator(
      Comparator<T> comparator) {
    return Ordering.from(comparator).reverse();
  }

  @Override
  public String toString() {

inferred type arguments.
Otherwise, type inference fails and the program is invalid.
</p>

<p>
Type inference uses the type relationships between pairs of types for inference:
For instance, a function argument must be <a href="#Assignability">assignable</a>
to its respective function parameter; this establishes a relationship between the
type of the argument and the type of the parameter.

      status->status = TF_TensorToTensor(input_tensors[i], &input_tensor);
      if (!status->status.ok()) return;
      input_tensors_vector[i] = &input_tensor;
    }
  }

  // Create an inference context with dummy values, which will be updated later.
  InferenceContext c(TF_GRAPH_DEF_VERSION, node_def, op_reg_data->op_def,
                     std::vector<ShapeHandle>(num_inputs), input_tensors_vector,

      SortedMap<K, ? extends V> sortedLeft = (SortedMap<K, ? extends V>) left;
      return difference(sortedLeft, right);
    }
    return difference(left, right, Equivalence.equals());
  }

  /**
   * Computes the difference between two maps. This difference is an immutable snapshot of the state
   * of the maps at the time this method is called. It will never change, even if the maps change at
   * a later time.
   *

*   Fixes an undefined behavior arising from reference binding to nullptr in
    shape inference
    ([CVE-2021-37676](https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2021-37676))
*   Fixes a division by 0 in most convolution operators
    ([CVE-2021-37675](https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2021-37675))
*   Fixes vulnerabilities arising from missing validation in shape inference for
    `Dequantize`

      SortedMap<K, ? extends V> sortedLeft = (SortedMap<K, ? extends V>) left;
      return difference(sortedLeft, right);
    }
    return difference(left, right, Equivalence.equals());
  }

  /**
   * Computes the difference between two maps. This difference is an immutable snapshot of the state
   * of the maps at the time this method is called. It will never change, even if the maps change at
   * a later time.
   *

  }

  TF_Output output_;

  // For LLVM style RTTI.
  static bool classof(const AbstractTensorHandle* ptr) {
    return ptr->getKind() == kGraph;
  }

 private:
  TF_Graph* graph_;  // For shape inference.
};

// GraphOperation wraps and populates a TF_OperationDescription.
class GraphOperation : public TracingOperation {
 public:
  explicit GraphOperation(TF_Graph* g) : TracingOperation(kGraph), g_(g) {}