- CRDs: For structural schemas, non-nullable null map fields will now be dropped and defaulted if a default is available. null items in the list will continue being preserved, and fail validation if not nullable. ([#95423](https://github.com/kubernetes/kubernetes/pull/95423), [@apelisse](https://github.com/apelisse)) [SIG API Machinery]

### Bug or Regression

- Do not fail sorting empty elements. ([#94666](https://github.com/kubernetes/kubernetes/pull/94666), [@soltysh](https://github.com/soltysh)) [SIG CLI]

- StatefulSet autodelete will respect controlling owners on PVC claims as described in https://github.com/kubernetes/enhancements/pull/4375 ([#126581](https://github.com/kubernetes/kubernetes/pull/126581), [@mattcary](https://github.com/mattcary)) [SIG Apps, Storage and Testing]

- In the client-side apply on create, defining the null value as "delete the key associated with this value". ([#125646](https://github.com/kubernetes/kubernetes/pull/125646), [@HirazawaUi](https://github.com/HirazawaUi)) [SIG API Machinery, CLI and Testing]

    val source = stream1.getSource()
    val buffer = Buffer()
    while (buffer.size != 1024L) source.read(buffer, 1024)
    stream1.close(ErrorCode.CANCEL, null)
    val frame1 = peer.takeFrame()
    assertThat(frame1.type).isEqualTo(Http2.TYPE_HEADERS)
    val frame2 = peer.takeFrame()
    assertThat(frame2.type).isEqualTo(Http2.TYPE_RST_STREAM)
    val frame3 = peer.takeFrame()

	// Avoid resyncing null versions created prior to enabling replication if target has a newer copy
	if opType == replication.ExistingObjectReplicationType &&
		oi1.ModTime.Unix() > oi2.LastModified.Unix() && oi1.VersionID == nullVersionID {
		return replicateNone
	}
	sz, _ := oi1.GetActualSize()

	// needs full replication
	if oi1.ETag != oi2.ETag ||
		oi1.VersionID != oi2.VersionID ||

			ObjectV: ObjectV{
				ObjectName: object,
				VersionID:  opts.VersionID,
			},
		}, *oi, opts, gerr)
		// Mutations of objects on versioning suspended buckets
		// affect its null version. Through opts below we select
		// the null version's remote object to delete if
		// transitioned.
		if gerr == nil {
			os.SetTransitionState(oi.TransitionedObject)
		}
		return dsc, nil
	})

- kube-apiserver: fixed a conflict error encountered attempting to delete a pod with `gracePeriodSeconds=0` and a resourceVersion precondition ([#85516](https://github.com/kubernetes/kubernetes/pull/85516), [@michaelgugino](https://github.com/michaelgugino))

		if dobj.ObjectName == "" {
			continue
		}

		if replicateDeletes && (dobj.DeleteMarkerReplicationStatus() == replication.Pending || dobj.VersionPurgeStatus() == Pending) {
			// copy so we can re-add null ID.
			dobj := dobj
			if isDirObject(dobj.ObjectName) && dobj.VersionID == "" {
				dobj.VersionID = nullVersionID
			}
			dv := DeletedObjectReplicationInfo{
				DeletedObject: dobj,
				Bucket:        bucket,

* Fix LoadBalancer rule checking so that no unexpected LoadBalancer updates are made ([#85990](https://github.com/kubernetes/kubernetes/pull/85990), [@feiskyer](https://github.com/feiskyer))