* which approach to use, <b>don't worry</b> too much about it; just pick something that seems
 * reasonable and it will be fine.
 *
 * <ul>
 *   <li>If checking whether the <i>caller</i> has violated your method or constructor's contract
 *       (such as by passing an invalid argument), use the utilities of the {@link Preconditions}
 *       class instead.

      thrownUnsafeFailure = unsafeFailure;
      // catch absolutely everything and fall through to our 'SafeAtomicHelper'
      // The access control checks that ARFU does means the caller class has to be AbstractFuture
      // instead of SafeAtomicHelper, so we annoyingly define these here
      try {
        helper =
            new SafeAtomicHelper(

 *
 * <p>{@code ByteSink} provides two kinds of methods:
 *
 * <ul>
 *   <li><b>Methods that return a stream:</b> These methods should return a <i>new</i>, independent
 *       instance each time they are called. The caller is responsible for ensuring that the
 *       returned stream is closed.
 *   <li><b>Convenience methods:</b> These are implementations of common operations that are

 *
 * <p>Generally, implementations don't support {@link #put} and {@link #putAll} because there is no
 * way to check an object at runtime to be an instance of a {@link TypeToken}. Instead, caller
 * should use the type safe {@link #putInstance}.
 *
 * <p>Also, if caller suppresses unchecked warnings and passes in an {@code Iterable<String>} for
 * type {@code Iterable<Integer>}, the map won't be able to detect and throw type error.
 *

   * elements than the collection), the element in the array immediately following the end of the
   * collection is set to {@code null}. This is useful in determining the length of the collection
   * <i>only</i> if the caller knows that the collection does not contain any null elements.
   *
   * <p>This method returns the elements in the order they are returned by the collection's
   * iterator.
   *

  }

  final Set<Throwable> getOrInitSeenExceptions() {
    /*
     * The initialization of seenExceptions has to be more complicated than we'd like. The simple
     * approach would be for each caller CAS it from null to a Set populated with its exception. But
     * there's another race: If the first thread fails with an exception and a second thread
     * immediately fails with the same exception:
     *

   * stream. When encoding it is "unflipped" (encoded bytes between 0 and position) and when
   * draining it is flipped (undrained bytes between position and limit).
   */
  private ByteBuffer byteBuffer;

  /** Whether we've finished reading the reader. */
  private boolean endOfInput;
  /** Whether we're copying encoded bytes to the caller's buffer. */
  private boolean draining;

 *
 * <p>{@code ByteSource} provides two kinds of methods:
 *
 * <ul>
 *   <li><b>Methods that return a stream:</b> These methods should return a <i>new</i>, independent
 *       instance each time they are called. The caller is responsible for ensuring that the
 *       returned stream is closed.
 *   <li><b>Convenience methods:</b> These are implementations of common operations that are

 *
 * <p>{@code CharSource} provides two kinds of methods:
 *
 * <ul>
 *   <li><b>Methods that return a reader:</b> These methods should return a <i>new</i>, independent
 *       instance each time they are called. The caller is responsible for ensuring that the
 *       returned reader is closed.
 *   <li><b>Convenience methods:</b> These are implementations of common operations that are

 * Iterator#remove} operations. This utility takes the brute-force approach of trying <i>all</i>
 * possible sequences of these operations, up to a given number of steps. So, if the caller
 * specifies to use <i>n</i> steps, a total of <i>3^n</i> tests are actually performed.
 *
 * <p>For instance, if <i>steps</i> is 5, one example sequence that will be tested is:
 *
 * <ol>
 *   <li>remove();