So we have a dedicated thread for every socket that just reads frames and dispatches them.

The reader thread must never run application-layer code. Otherwise one slow stream can hold up the entire connection.

  /*
   * TODO: consider adding MultisetFeatures.SUPPORTS_SET_COUNT. Currently we
   * assume that using setCount() to increase the count is permitted iff add()
   * is permitted and similarly for decrease/remove(). We assume that a
   * setCount() no-op is permitted if either add() or remove() is permitted,
   * though we also allow it to "succeed" if neither is permitted.
   */

You can build your own connection spec with a custom set of TLS versions and cipher suites. For example, this configuration is limited to three highly-regarded cipher suites. Its drawback is that it requires Android 5.0+ and a similarly current webserver.

```java
ConnectionSpec spec = new ConnectionSpec.Builder(ConnectionSpec.MODERN_TLS)
    .tlsVersions(TlsVersion.TLS_1_2)
    .cipherSuites(

  //   timeout is small we shouldn't park(). This needs to be traded off with the cpu overhead of
  //   spinning, so we use SPIN_THRESHOLD_NANOS which is what AbstractQueuedSynchronizer uses for
  //   similar purposes.
  // * We want to behave reasonably for timeouts of 0
  // * We are more responsive to completion than timeouts. This is because parkNanos depends on

 *       MoreExecutors#listeningDecorator(java.util.concurrent.ExecutorService)
 *       MoreExecutors.listeningDecorator}.
 *   <li>If you manually call {@link java.util.concurrent.FutureTask#set} or a similar method,
 *       create a {@link SettableFuture} instead. (If your needs are more complex, you may prefer
 *       {@link AbstractFuture}.)
 * </ul>
 *

 *       MoreExecutors#listeningDecorator(java.util.concurrent.ExecutorService)
 *       MoreExecutors.listeningDecorator}.
 *   <li>If you manually call {@link java.util.concurrent.FutureTask#set} or a similar method,
 *       create a {@link SettableFuture} instead. (If your needs are more complex, you may prefer
 *       {@link AbstractFuture}.)
 * </ul>
 *

  }

  /**
   * A special purpose queue/executor that dispatches listener events serially on a configured
   * executor. Each event can be added and dispatched as separate phases.
   *
   * <p>This class is very similar to {@link SequentialExecutor} with the exception that events can
   * be added without necessarily executing immediately.
   */
  private static final class PerListenerQueue<L> implements Runnable {
    final L listener;

            throw new IOException("Connection not established");
        }

        try {
            // In real implementation, this would perform RDMA write:
            // Similar to rdmaRead but with IBV_WR_RDMA_WRITE opcode

            log.debug("DiSNI RDMA write completed, {} bytes", length);

        } catch (Exception e) {
            state = RdmaConnectionState.ERROR;

    @Nonnull
    Optional<Boolean> ignoreTransitiveRepositories();

    /**
     * Specifies "@file"-like file, to load up command line from. It may contain goals as well. Format is one parameter
     * per line (similar to {@code maven.conf}) and {@code '#'} (hash) marked comment lines are allowed. Goals, if
     * present, are appended, to those specified on CLI input, if any.
     */
    Optional<String> atFile();

    /**

	if err != nil {
		return "", err
	}
	return trimLeadingSlash(ep), nil
}

// unescapePath is similar to unescapeGeneric but for specifically
// path unescaping.
func unescapePath(p string) (string, error) {
	return unescapeGeneric(p, url.PathUnescape)
}

// similar to unescapeGeneric but never returns any error if the unescaping
// fails, returns the input as is in such occasion, not meant to be