// Might also explore collocating all of [hash, next, predecessor, successor] fields of an
  // entry in a *single* long[], though that reduces the maximum size of the set by a factor of 2

  /**
   * Pointer to the predecessor of an entry in insertion order. ENDPOINT indicates a node is the
   * first node in insertion order; all values at indices ≥ {@link #size()} are UNSET.
   */

      super(key, value);
      this.keyHash = keyHash;
      this.valueHash = valueHash;
    }
  }

  private static final double LOAD_FACTOR = 1.0;

  /*
   * The following two arrays may *contain* nulls, but they are never *themselves* null: Even though
   * they are not initialized inline in the constructor, they are initialized from init(), which the

        Serialization.getFieldSetter(ConcurrentHashMultiset.class, "countMap");
  }

  /**
   * Creates a new, empty {@code ConcurrentHashMultiset} using the default initial capacity, load
   * factor, and concurrency settings.
   */
  public static <E> ConcurrentHashMultiset<E> create() {
    // TODO(schmoe): provide a way to use this class with other (possibly arbitrary)

         * negative so it with NT_STATUS_SUCCESS (0) the binary search will not be
         * performed properly. The effect is that the code at index 1 is never found
         * (NT_STATUS_UNSUCCESSFUL). So here we factor out NT_STATUS_SUCCESS
         * as a special case (which it is).
         */
        if (errcode == 0) {
            return "NT_STATUS_SUCCESS";
        }
        if(( errcode & 0xC0000000 ) == 0xC0000000 ) {

              knownRunStart += maxRunBeforeFallback;
            } else {
              knownRunStart++; // the only case in which maxRunEnd doesn't increase by mRBF
              // happens about f * (1-f) for f = DESIRED_LOAD_FACTOR, so around 21% of the time
            }
            knownRunEnd = knownRunStart;
          } else {
            for (int j = knownRunStart + maxRunBeforeFallback - 1; j >= knownRunEnd; j--) {

- Already compressed objects are not fit for compression since they do not have compressible patterns.
Such objects do not produce efficient [`LZ compression`](https://en.wikipedia.org/wiki/LZ77_and_LZ78)
which is a fitness factor for a lossless data compression.

Pre-compressed input typically compresses in excess of 2GiB/s per core,
so performance impact should be minimal even if precompressed data is re-compressed.

    implements ClassToInstanceMap<B>, Serializable {

  /**
   * Returns a new {@code MutableClassToInstanceMap} instance backed by a {@link HashMap} using the
   * default initial capacity and load factor.
   */
  public static <B extends @Nullable Object> MutableClassToInstanceMap<B> create() {
    return new MutableClassToInstanceMap<>(new HashMap<Class<? extends @NonNull B>, B>());
  }

  /**

        Serialization.getFieldSetter(ConcurrentHashMultiset.class, "countMap");
  }

  /**
   * Creates a new, empty {@code ConcurrentHashMultiset} using the default initial capacity, load
   * factor, and concurrency settings.
   */
  public static <E> ConcurrentHashMultiset<E> create() {
    // TODO(schmoe): provide a way to use this class with other (possibly arbitrary)

  // Might also explore collocating all of [hash, next, predecessor, successor] fields of an
  // entry in a *single* long[], though that reduces the maximum size of the set by a factor of 2

  /**
   * Pointer to the predecessor of an entry in insertion order. ENDPOINT indicates a node is the
   * first node in insertion order; all values at indices ≥ {@link #size()} are UNSET.
   */

- In this algorithm, we also make sure that we spread the drives out evenly. MinIO server expands ellipses passed as arguments. Here is a sample expansion to demonstrate the process.