*
       * My impression is that an AtomicReferenceFieldUpdater in a static field is similarly fast to
       * Unsafe on modern JVMs (if perhaps not quite as fast as VarHandle?). However, I'm not sure
       * exactly what we've benchmarked, and we certainly haven't benchmarked as far back as JDK 8.
       * (We also haven't benchmarked under Android. We continue to use UnsafeAtomicHelper there so

   * identical inputs: This sounds like it ought to be a problem here, since the goals of this class
   * include testing that {@code equals()} is reflexive and is tolerant of {@code null}. However,
   * there's no problem. The reason: {@link EqualsTester} tests {@code null} and identical inputs
   * directly against {@code equals()} rather than through the {@code Equivalence}.
   */

 * {@code asMap} has a few subtleties. As long as the set of keys remains unchanged, adding or
 * removing mappings does not affect the key iteration order. However, if you remove all values
 * associated with a key and then add the key back to the multimap, that key will come last in the
 * key iteration order.
 *

   * the backport does. For now, we're keeping the class declaration in *both* flavors so that both
   * flavors can read old data or data from the other flavor. However, we strongly discourage
   * relying on this, as we have made incompatible changes to serialized forms in the past and
   * expect to do so again, as discussed in https://github.com/google/guava#important-warnings.
   */

   * add the suppression to the whole method (which turns off checking for a large section of code),
   * 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

 * interface. It does not permit null keys or values.
 *
 * <p><b>Note:</b> by default, the returned map uses equality comparisons (the {@link Object#equals
 * equals} method) to determine equality for keys or values. However, if {@link #weakKeys} was
 * specified, the map uses identity ({@code ==}) comparisons instead for keys. Likewise, if {@link
 * #weakValues} was specified, the map uses identity comparisons for values.
 *

   * bug</b> as described in the deprecation text.
   *
   * <p>The C++ equivalent is the MurmurHash3_x86_32 function (Murmur3A), which however does not
   * have the bug.
   *
   * @deprecated This implementation produces incorrect hash values from the {@link
   *     HashFunction#hashString} method if the string contains non-BMP characters. Use {@link

   * @param length the length of the new array
   */
  /*
   * The new array contains nulls, even if the old array did not. If we wanted to be accurate, we
   * would declare a return type of `@Nullable T[]`. However, we've decided not to think too hard
   * about arrays for now, as they're a mess. (We previously discussed this in the review of
   * ObjectArrays, which is the main caller of this method.)
   */

        .testEquals();
  }

  /*
   * We use large numbers to avoid the integer cache. Normally, we'd accomplish that merely by using
   * `new Integer` (as we do) instead of `Integer.valueOf`. However, under J2KT, `new Integer`
   * gets translated back to `Integer.valueOf` because that is the only thing J2KT can support. And
   * anyway, it's nice to avoid `Integer.valueOf` because the Android toolchain optimizes multiple

   * identical inputs: This sounds like it ought to be a problem here, since the goals of this class
   * include testing that {@code equals()} is reflexive and is tolerant of {@code null}. However,
   * there's no problem. The reason: {@link EqualsTester} tests {@code null} and identical inputs
   * directly against {@code equals()} rather than through the {@code Equivalence}.
   */