| ((source[offset + 1] & 0xFF) << 8)
        | ((source[offset + 2] & 0xFF) << 16)
        | ((source[offset + 3] & 0xFF) << 24);
  }

  /**
   * Indicates that the load and store operations will be very efficient because of use of VarHandle
   * or Unsafe. May be useful for calling code to fall back on an alternative implementation that is
   * slower than those implementations but faster than the pure-Java mask-and-shift.
   */

        synchronized (lock) {
          if (nanos == expirationNanos) { // recheck for lost race
            T t = delegate.get();
            value = t;
            nanos = now + durationNanos;
            // In the very unlikely event that nanos is 0, set it to 1;
            // no one will notice 1 ns of tardiness.
            expirationNanos = (nanos == 0) ? 1 : nanos;
            return t;
          }
        }
      }

        synchronized (lock) {
          if (nanos == expirationNanos) { // recheck for lost race
            T t = delegate.get();
            value = t;
            nanos = now + durationNanos;
            // In the very unlikely event that nanos is 0, set it to 1;
            // no one will notice 1 ns of tardiness.
            expirationNanos = (nanos == 0) ? 1 : nanos;
            return t;
          }
        }
      }

 *       hand, is multiple-use, and does implement {@link Iterable}.
 *   <li>Streams offer many features not found here, including {@code min/max}, {@code distinct},
 *       {@code reduce}, {@code sorted}, the very powerful {@code collect}, and built-in support for
 *       parallelizing stream operations.
 *   <li>{@code FluentIterable} contains several features not available on {@code Stream}, which are

 *       hand, is multiple-use, and does implement {@link Iterable}.
 *   <li>Streams offer many features not found here, including {@code min/max}, {@code distinct},
 *       {@code reduce}, {@code sorted}, the very powerful {@code collect}, and built-in support for
 *       parallelizing stream operations.
 *   <li>{@code FluentIterable} contains several features not available on {@code Stream}, which are

      * If you use the very rarely-used function `tf.keras.mixed_precision.experimental.get_layer_policy`:
          * Replace `tf.keras.mixed_precision.experimental.get_layer_policy(layer)` with `layer.dtype_policy`.

## 0.120.0 (2025-10-23)

There are no major nor breaking changes in this release. ☕️

The internal reference documentation now uses `annotated_doc.Doc` instead of `typing_extensions.Doc`, this adds a new (very small) dependency on [`annotated-doc`](https://github.com/fastapi/annotated-doc), a package made just to provide that `Doc` documentation utility class.

   * map.
   *
   * <p>The returned map isn't threadsafe or serializable, even if {@code unfiltered} is.
   *
   * <p>Many of the filtered map's methods, such as {@code size()}, iterate across every key/value
   * mapping in the underlying map and determine which satisfy the filter. When a live view is
   * <i>not</i> needed, it may be faster to copy the filtered map and use the copy.
   *

   * map.
   *
   * <p>The returned map isn't threadsafe or serializable, even if {@code unfiltered} is.
   *
   * <p>Many of the filtered map's methods, such as {@code size()}, iterate across every key/value
   * mapping in the underlying map and determine which satisfy the filter. When a live view is
   * <i>not</i> needed, it may be faster to copy the filtered map and use the copy.
   *

   *       Executor)} instead, with a function that returns the next value directly.
   *   <li>Call {@link DeferredCloser#eventuallyClose(Object, Executor) closer.eventuallyClose()}
   *       for every closeable object this step creates in order to capture it for later closing.
   *   <li>Return a {@code ClosingFuture}. To turn a {@link ListenableFuture} into a {@code
   *       ClosingFuture} call {@link #from(ListenableFuture)}.