one hand and a piece of bread-and-butter in the other.  `I beg
pardon, your Majesty,' he began, `for bringing these in:  but I
hadn't quite finished my tea when I was sent for.'

  `You ought to have finished,' said the King.  `When did you
begin?'

  The Hatter looked at the March Hare, who had followed him into
the court, arm-in-arm with the Dormouse.  `Fourteenth of March, I

* Fix panic in kubectl cp command ([#75037](https://github.com/kubernetes/kubernetes/pull/75037), [@soltysh](https://github.com/soltysh))
* Fixes an issue with missing apiVersion/kind in object data sent to admission webhooks ([#74448](https://github.com/kubernetes/kubernetes/pull/74448), [@liggitt](https://github.com/liggitt))

          .stream()
          .filter { event: CallEvent? -> event is ConnectStart }
          .count()
      assertThat(connectCount).isEqualTo(1)
    }
  }

  /** Test which headers are sent unencrypted to the HTTP proxy.  */
  @Test
  fun proxyConnectOmitsApplicationHeaders() {
    server.useHttps(handshakeCertificates.sslSocketFactory())
    server.enqueue(
      MockResponse
        .Builder()

calls  with checking of ResourceVersion in the Watch call. This closes a race condition where creating an object directly after an informer cache has synced (= List call completed) and before the Watch call completed would cause that object to not be sent to the informer. A visible side-effect of adding that support is that List meta data contains a ResourceVersion (starting at "1" for the empty set, incremented by one  for each add/update) and that Watch may return objects where it previously didn't....

- Bookmark events are now sent immediately after all items in the watchCache store have been processed, improving consistency in client behavior. ([#127012](https://github.com/kubernetes/kubernetes/pull/127012), [@Chaunceyctx](https://github.com/Chaunceyctx))

- Fixed an issue with missing `apiVersion/kind` in object data sent to admission webhooks ([#74448](https://github.com/kubernetes/kubernetes/pull/74448), [@liggitt](https://github.com/liggitt))

- The audit event sourceIPs list will now always end with the IP that sent the request directly to the API server. ([#87167](https://github.com/kubernetes/kubernetes/pull/87167), [@tallclair](https://github.com/tallclair)) [SIG API Machinery and Auth]

* fix apps DeploymentSpec conversion issue ([#49719](https://github.com/kubernetes/kubernetes/pull/49719), [@dixudx](https://github.com/dixudx))
* fluentd-gcp addon: Fix a bug in the event-exporter, when repeated events were not sent to Stackdriver. ([#50511](https://github.com/kubernetes/kubernetes/pull/50511), [@crassirostris](https://github.com/crassirostris))

- Fixed a kubelet issue that could result in invalid pod status updates to be sent to the api-server where pods would be reported in a terminal phase but also report a ready condition of true in some cases. ([#110481](https://github.com/kubernetes/kubernetes/pull/110481), [@bobbypage](https://github.com/bobbypage)) [SIG Node and...

* On GCE, the apiserver audit log (`/var/log/kube-apiserver-audit.log`) will be sent through fluentd if enabled. It will go to the same place as `kube-apiserver.log`, but tagged as its own stream. ([#41360](https://github.com/kubernetes/kubernetes/pull/41360), [@enisoc](https://github.com/enisoc))