* constant time operations. Expected in the hashtable sense (depends on the hash function doing a
 * good job of distributing the elements to the buckets to a distribution not far from uniform), and
 * amortized since some operations can trigger a hash table resize.
 *
 * <p>Unlike {@code java.util.HashSet}, iteration is only proportional to the actual {@code size()},

O importante aqui de se lembrar é que **o primeiro *type parameter*** que você informar ao `Annotated` é o **tipo de fato**. O resto é apenas metadado para outras ferramentas.

Por hora, você precisa apenas saber que o `Annotated` existe, e que ele é Python padrão. 😎

* NodeController waits for informer sync before doing anything ([#34809](https://github.com/kubernetes/kubernetes/pull/34809), [@gmarek](https://github.com/gmarek))

}

// Retrieves the external IP of the ingress-gateway for the hosts file additions
func createHosts(kubeClient kube.CLIClient, istioNamespace, ingressIP, dir string, revision string) error {
	// try to infer the ingress IP if the provided one is invalid
	if agent.ValidateIPAddress(ingressIP) != nil {
		p := strings.Split(ingressSvc, ".")
		ingressNs := istioNamespace
		if len(p) == 2 {
			ingressSvc = p[0]

  // Kind of the referent.
  // More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
  optional string kind = 1;

  // Name of the referent.
  // More info: https://kubernetes.io/docs/concepts/overview/working-with-objects/names#names
  optional string name = 3;

  // UID of the referent.

const (
	opLt    = "<"
	opLte   = "<="
	opGt    = ">"
	opGte   = ">="
	opEq    = "="
	opIneq  = "!="
	opIs    = "IS"
	opIsNot = "ISNOT"
)

// InferBytesType will attempt to infer the data type of bytes.
// Will fail if value type is not bytes or it would result in invalid utf8.
// ORDER: int, float, bool, JSON (object or array), timestamp, string

  For an executable, the required form of the "work that uses the
Library" must include any data and utility programs needed for
reproducing the executable from it.  However, as a special exception,
the materials to be distributed need not include anything that is
normally distributed (in either source or binary form) with the major
components (compiler, kernel, and so on) of the operating system on

 * form}), the memory requirement is 8*N bytes for the copy of the dataset plus an overhead which is
 * independent of N (but depends on the quantiles being computed). When calling {@link
 * ScaleAndIndex#computeInPlace computeInPlace} (in {@linkplain ScaleAndIndexes#computeInPlace
 * either form}), only the overhead is required. The number of object allocations is independent of

   * table is internally partitioned to try to permit the indicated number of concurrent updates
   * without contention. Because assignment of entries to these partitions is not necessarily
   * uniform, the actual concurrency observed may vary. Ideally, you should choose a value to
   * accommodate as many threads as will ever concurrently modify the table. Using a significantly

* Fixed an issue with informers missing an `Added` event if a recently deleted object was immediately recreated at the same time the informer dropped a watch and relisted. ([#83911](https://github.com/kubernetes/kubernetes/pull/83911), [@matte21](https://github.com/matte21))