* All registered subscribers, indexed by event type.
   *
   * <p>The {@link CopyOnWriteArraySet} values make it easy and relatively lightweight to get an
   * immutable snapshot of all current subscribers to an event without any locking.
   */
  private final ConcurrentMap<Class<?>, CopyOnWriteArraySet<Subscriber>> subscribers =
      Maps.newConcurrentMap();

  /** The event bus this registry belongs to. */
  @Weak private final EventBus bus;

* Fix polarity of NodePort logic to avoid leaked ports ([#43259](https://github.com/kubernetes/kubernetes/pull/43259), [@thockin](https://github.com/thockin))
* Upgrade golang versions to 1.7.5 ([#41771](https://github.com/kubernetes/kubernetes/pull/41771), [@cblecker](https://github.com/cblecker))



# v1.5.6

    // way for it to return a false negative would be for the target value to jump around in the map
    // such that none of the subsequent iterations observed it, despite the fact that at every point
    // in time it was present somewhere int the map. This becomes increasingly unlikely as
    // CONTAINS_VALUE_RETRIES increases, though without locking it is theoretically possible.

 * multiple routes to be attempted simultaneously.
 *
 * Events and sequences of events may be repeated for retries and follow-ups.
 *
 * All event methods must execute fast, without external locking, cannot throw exceptions, attempt
 * to mutate the event parameters, or be re-entrant back into the client. Any IO - writing to files
 * or network should be done asynchronously.
 */
abstract class EventListener {
  /**

        // Should not throw NullPointerException
        try {
            // Note: We can't easily test connect() fully without mocking createClient()
            // This test verifies the null check logic indirectly
            final String[] indices = fesenClient.targetIndices != null ? fesenClient.targetIndices : new String[0];

            unicodeString.buffer = new short[] { 100, 200, 300 };

            // Create a real NdrBuffer mock that can handle the deferred field access
            // But only test the non-deferred parts to avoid field mocking issues

            // When: Trying to encode (may throw exception due to deferred buffer)
            // Then: Should at least call the basic encoding methods before hitting deferred

- The resource/v1alpha3.ResourceSliceList filed which should have been named "metadata" but was instead named "listMeta" is now properly "metadata". ([#126749](https://github.com/kubernetes/kubernetes/pull/126749), [@thockin](https://github.com/thockin)) [SIG API Machinery]

* Change merge key for VolumeMount to mountPath ([#35071](https://github.com/kubernetes/kubernetes/pull/35071), [@thockin](https://github.com/thockin))
* Turned-off etcd listening on public ports as potentially insecure. Removed ([#35192](https://github.com/kubernetes/kubernetes/pull/35192), [@jszczepkowski](https://github.com/jszczepkowski))

     * for (Lock lock : locks) {
     *   lock.unlock();
     * }
     *
     * If we only held the int[] stripes, translating it on the fly to L's, the original locks might
     * be garbage collected after locking them, ending up in a huge mess.
     */
    @SuppressWarnings("unchecked") // we carefully replaced all keys with their respective L's
    List<L> asStripes = (List<L>) result;

* Scale kube-proxy conntrack limits by cores (new default behavior) ([#28876](https://github.com/kubernetes/kubernetes/pull/28876), [@thockin](https://github.com/thockin))
* Don't recreate lb cloud resources on kcm restart ([#29082](https://github.com/kubernetes/kubernetes/pull/29082), [@bprashanth](https://github.com/bprashanth))

## Known Issues