* executor. The listener will have the corresponding transition method called whenever the
   * service changes state. The listener will not have previous state changes replayed, so it is
   * suggested that listeners are added before the service starts.
   *
   * <p>{@code addListener} guarantees execution ordering across calls to a given listener but not

   * That is, the first call to {@link #nextIndex()} will return {@code position}, and the first
   * call to {@link #next()} will return the element at that index, if available. Calls to {@link
   * #previous()} can retrieve the preceding {@code position} elements.
   *
   * @throws IndexOutOfBoundsException if {@code position} is negative or is greater than {@code
   *     size}

    Map<K, V> delegate = Maps.newLinkedHashMap();
    for (Entry<? extends K, ? extends V> entry : entries) {
      K key = checkNotNull(entry.getKey());
      V previous = delegate.put(key, checkNotNull(entry.getValue()));
      if (throwIfDuplicateKeys && previous != null) {
        throw new IllegalArgumentException("duplicate key: " + key);
      }
    }
    this.delegate = Collections.unmodifiableMap(delegate);
  }

    /**
     * The elements to be returned by future calls to {@code previous()}, with the first at the top
     * of the stack.
     */
    final Stack<E> previousElements = new Stack<>();

    /**
     * {@link #nextElements} if {@code next()} was called more recently then {@code previous},
     * {@link #previousElements} if the reverse is true, or -- overriding both of these -- {@code

      } else {
        // we need to update a pointer in an entry
        int previous;
        long entry;
        do {
          previous = lastNext;
          lastNext = getNext(entry = entries[lastNext]);
        } while (lastNext != srcIndex);
        // here, entries[previous] points to the old entry location; update it
        entries[previous] = swapNext(entry, dstIndex);
      }
    } else {
      keys[dstIndex] = null;

# limitations under the License.
# ==============================================================================
# This script uploads all staged artifacts from all previous builds in the same
# job chain to GCS and PyPI.
source "${BASH_SOURCE%/*}/utilities/setup.sh"

# Calculate the version number for choosing the final directory name. This adds

| xlVersionMinor | uint16         | Minor xl-meta version.
| xlMetaV2       | msgp bin array | Bin array with serialized metadata
| crc            | msgp uint      | Lower 32 bits of 64 bit xxhash of previous array contents (v1.2+ only)
| inline data    | binary         | Inline data if any, see Inline Data section for encoding.  
| [EOF] | |

## v1.0-v1.2 Versions

/**
 * Serializes execution of tasks, somewhat like an "asynchronous {@code synchronized} block." Each
 * {@linkplain #submit enqueued} callable will not be submitted to its associated executor until the
 * previous callable has returned -- and, if the previous callable was an {@link AsyncCallable}, not
 * until the {@code Future} it returned is {@linkplain Future#isDone done} (successful, failed, or
 * cancelled).
 *

  }

  /**
   * Atomically sets the element at position {@code i} to the given value and returns the old value.
   *
   * @param i the index
   * @param newValue the new value
   * @return the previous value
   */
  public final double getAndSet(int i, double newValue) {
    long next = doubleToRawLongBits(newValue);
    return longBitsToDouble(longs.getAndSet(i, next));
  }

  /**

     *
     * @param config the configuration to use
     * @param sid the search ID from a previous FindFirst2 response
     * @param resumeKey the resume key for continuing the search
     * @param filename the last filename from the previous response
     * @param batchCount the number of entries to return
     * @param batchSize the maximum size of the response buffer
     */