until we've pushed its bytes onto the socket. Otherwise, if the write fails we are unable to deliver its IOException to the application. We would have told the application layer that the write succeeded, but it didn't!

The application-layer can also do blocking reads. If the application asks to read and there's nothing available, we need to hold that thread until either the bytes arrive, the stream is closed, or a timeout elapses. If we get bytes but there's nobody asking for them, we buffer...

case{text-transform:uppercase!important}.text-capitalize{text-transform:capitalize!important}.font-weight-light{font-weight:300!important}.font-weight-lighter{font-weight:lighter!important}.font-weight-normal{font-weight:400!important}.font-weight-bold{font-weight:700!important}.font-weight-bolder{font-weight:bolder!important}.font-italic{font-style:italic!important}.text-white{color:#fff!important}.text-primary{color:#007bff!important}a.text-primary:focus,a.text-primary:hover{color:#0056b3!impo...

1. Shortly after, Kubernetes will start terminating `ztunnel-old`. It does this initially by sending a `SIGTERM`. `ztunnel-old` will catch this, and start "draining".
1. Immediately upon starting a drain, `ztunnel-old` will close its listeners. Now only `ztunnel-new` is listening. Critically, at all times there was at least one ztunnel listening.
1. While `ztunnel-old` will not accept *new* connections, it will continue processing existing connections.

			// Unlike the other cases, we actually want to use the "old" event for terminated job pods
			// - kubernetes will (weirdly) issue a new status to the pod with no IP on termination, meaning
			// our check of `pod.status` will fail for (some) termination events.
			//
			// We will get subsequent events that append a new status with the IP put back, but it's simpler
			// and safer to just check the old pod status for the IP.

    /** Constructor for use by subclasses. */
    protected SleepingStopwatch() {}

    /*
     * We always hold the mutex when calling this. TODO(cpovirk): Is that important? Perhaps we need
     * to guarantee that each call to reserveEarliestAvailable, etc. sees a value >= the previous?
     * Also, is it OK that we don't hold the mutex when sleeping?
     */
    protected abstract long readMicros();

  abstract static class OldAbstractFuture<V> implements ListenableFuture<V> {

    /** Synchronization control for AbstractFutures. */
    private final Sync<V> sync = new Sync<V>();

    // The execution list to hold our executors.
    private final ExecutionList executionList = new ExecutionList();

    /** Constructor for use by subclasses. */
    protected OldAbstractFuture() {}

    /*

italics\n\n.font-weight-light   { font-weight: $font-weight-light !important; }\n.font-weight-lighter { font-weight: $font-weight-lighter !important; }\n.font-weight-normal  { font-weight: $font-weight-normal !important; }\n.font-weight-bold    { font-weight: $font-weight-bold !important; }\n.font-weight-bolder  { font-weight: $font-weight-bolder !important; }\n.font-italic         { font-style: italic !important; }\n\n// Contextual colors\n\n.text-white { color: $white !important; }\n\n@each $color, $value...

and kept doubling itself up and straightening itself out again,
so that altogether, for the first minute or two, it was as much
as she could do to hold it.

  As soon as she had made out the proper way of nursing it,
(which was to twist it up into a sort of knot, and then keep
tight hold of its right ear and left foot, so as to prevent its
undoing itself,) she carried it out into the open air.  `IF I

  /**
   * Creates a {@code CompactHashSet} instance, with a high enough "initial capacity" that it
   * <i>should</i> hold {@code expectedSize} elements without growth.
   *
   * @param expectedSize the number of elements you expect to add to the returned set
   * @return a new, empty {@code CompactHashSet} with enough capacity to hold {@code expectedSize}
   *     elements without resizing

	"istio.io/istio/pkg/zdsapi"
)

var ErrPodNotFound = errors.New("netns not provided, but is needed as pod is not in cache")

type PodNetnsCache interface {
	ReadCurrentPodSnapshot() map[string]WorkloadInfo
}

// Hold a cache of node local pods with their netns
// if we don't know the netns, the pod will still be here with a nil netns.
type podNetnsCache struct {
	openNetns func(nspath string) (NetnsCloser, error)