### EventListener.Factory

In the preceding example we used a field, `callStartNanos`, to track the elapsed time of each event. This is handy, but it won’t work if multiple calls are executing concurrently. To accommodate this, use a `Factory` to create a new `EventListener` instance for each `Call`. This allows each listener to keep call-specific state.

        }

        if (connectResult.isSuccess) {
          // We have a connected TCP connection. Cancel and defer the racing connects that all lost.
          cancelInFlightConnects()

          // Finish connecting. We won't have to if the winner is from the connection pool.
          if (!connectResult.plan.isReady) {
            connectResult = connectResult.plan.connectTlsEtc()
          }

          if (connectResult.isSuccess) {

### Conditional Cache Hit
 
When cache flags require checking the cache results are still valid an early cacheConditionalHit event is
received followed by a cache hit or miss.  Critically in the cache hit scenario the server won’t send the response body.

The response will have non-null `cacheResponse` and `networkResponse`. The cacheResponse will be used as the top level
response only if the response code is HTTP/1.1 304 Not Modified.
 
 - CallStart

 * connections should expect repetition of stream IDs.
 *
 * Return true to request cancellation of a pushed stream.  Note that this does not guarantee
 * future frames won't arrive on the stream ID.
 */
interface PushObserver {
  /**
   * Describes the request that the server intends to push a response for.
   *
   * @param streamId server-initiated stream ID: an even number.

    assertThat(statusLine.protocol).isEqualTo(Protocol.HTTP_1_1)
    assertThat(statusLine.code).isEqualTo(code)
  }

  /**
   * This is not defined in the protocol but some servers won't add the leading empty space when the
   * message is empty. http://www.w3.org/Protocols/rfc2616/rfc2616-sec6.html#sec6.1
   */
  @Test
  fun emptyMessageAndNoLeadingSpace() {
    val version = 1
    val code = 503

import okhttp3.Handshake;
import okhttp3.HttpUrl;
import okhttp3.OkHttpClient;
import okhttp3.Protocol;
import okhttp3.Request;
import okhttp3.Response;

/**
 * This prints events for a single in-flight call. It won't work for multiple concurrent calls
 * because we don't know what callStartNanos refers to.
 */
public final class PrintEventsNonConcurrent {
  private final OkHttpClient client = new OkHttpClient.Builder()

 * this class separates _allocations_ from _streams_. An allocation is created by a call, used for
 * one or more streams, and then released. An allocated connection won't be stolen by other calls
 * while a redirect or authorization challenge is being handled.
 *
 * When the maximum concurrent streams limit is reduced, some allocations will be rescinded.

 *
 * ## Web Socket Lifecycle
 *
 * Upon normal operation each web socket progresses through a sequence of states:
 *
 *  * **Connecting:** the initial state of each web socket. Messages may be enqueued but they won't
 *    be transmitted until the web socket is open.
 *
 *  * **Open:** the web socket has been accepted by the remote peer and is fully operational.
 *    Messages in either direction are enqueued for immediate transmission.