*     Guava.
   */
  @Override
  @CheckForNull
  public final V getOrDefault(@CheckForNull Object key, @CheckForNull V defaultValue) {
    /*
     * Even though it's weird to pass a defaultValue that is null, some callers do so. Those who
     * pass a literal "null" should probably just use `get`, but I would expect other callers to
     * pass an expression that *might* be null. This could happen with:

   *     Guava.
   */
  @Override
  @CheckForNull
  public final V getOrDefault(@CheckForNull Object key, @CheckForNull V defaultValue) {
    /*
     * Even though it's weird to pass a defaultValue that is null, some callers do so. Those who
     * pass a literal "null" should probably just use `get`, but I would expect other callers to

     *
     * OkHttpClient client = new OkHttpClient.Builder()
     *     .sslSocketFactory(sslSocketFactory, trustManager)
     *     .build();
     * ```
     *
     * ## TrustManagers on Android are Weird!
     *
     * Trust managers targeting Android must also define a method that has this signature:
     *
     * ```java
     *    @SuppressWarnings("unused")

font-weight: 400;\n  line-height: 1.5;\n  text-align: left;\n  text-align: start;\n  text-decoration: none;\n  text-shadow: none;\n  text-transform: none;\n  letter-spacing: normal;\n  word-break: normal;\n  word-spacing: normal;\n  white-space: normal;\n  line-break: auto;\n  font-size: 0.875rem;\n  word-wrap: break-word;\n  opacity: 0;\n}\n\n.tooltip.show {\n  opacity: 0.9;\n}\n\n.tooltip .arrow {\n  position: absolute;\n  display: block;\n  width: 0.8rem;\n  height: 0.4rem;\n}\n\n.tooltip .arrow::before...

          return null;
        }
      }
      return topMetaIterator;
    }

    @Override
    public boolean hasNext() {
      while (!checkNotNull(iterator).hasNext()) {
        // this weird checkNotNull positioning appears required by our tests, which expect
        // both hasNext and next to throw NPE if an input iterator is null.

        topMetaIterator = getTopMetaIterator();

    @Override
    public int hashCode() {
      // Warning: this is broken if size() == 0, so it is critical that we
      // substitute an empty ImmutableSet to the user in place of this

      // It's a weird formula, but tests prove it works.
      int adjust = size() - 1;
      for (int i = 0; i < axes.size(); i++) {
        adjust *= 31;
        adjust = ~~adjust;

but only worthwhile if it's likely the instruction will be used again.
Instead, for simple one-off cases, it's possible to use the <code>BYTE</code>
and <code>WORD</code> directives
to lay down explicit data into the instruction stream within a <code>TEXT</code>.
Here's how the 386 runtime defines the 64-bit atomic load function.
</p>

<pre>

    direct clients to retry rather than failing permanently.
 *  New: Expose the framed protocol in `Connection`. Previously this would
    return the application-layer protocol (HTTP/1.1 or HTTP/1.0); now it always
    returns the wire-layer protocol (HTTP/2, SPDY/3.1, or HTTP/1.1).
 *  Fix: Permit the trusted CA root to be pinned by `CertificatePinner`.
 *  Fix: Silently ignore unknown HTTP/2 settings. Previously this would cause

Emoji","Segoe UI Emoji","Segoe UI Symbol";font-style:normal;font-weight:400;line-height:1.5;text-align:left;text-align:start;text-decoration:none;text-shadow:none;text-transform:none;letter-spacing:normal;word-break:normal;word-spacing:normal;white-space:normal;line-break:auto;font-size:.875rem;word-wrap:break-word;opacity:0}.tooltip.show{opacity:.9}.tooltip .arrow{position:absolute;display:block;width:.8rem;height:.4rem}.tooltip .arrow::before{position:absolute;content:"";border-color:transparent;border-s...

   * bits.
   *
   * @param h hash code
   */
  static int rehash(int h) {
    // Spread bits to regularize both segment and index locations,
    // using variant of single-word Wang/Jenkins hash.
    // TODO(kevinb): use Hashing/move this to Hashing?
    h += (h << 15) ^ 0xffffcd7d;
    h ^= (h >>> 10);
    h += (h << 3);
    h ^= (h >>> 6);
    h += (h << 2) + (h << 14);