low = mid;
      } else {
        return mid; // allRequired[mid] = centerFloor, so we can't get closer than that
      }
    }

    // Now pick the closest of the two candidates. Note that there is no rounding here.
    if (from + to - allRequired[low] - allRequired[high] > 0) {
      return high;
    } else {
      return low;
    }
  }

- When a PreFilterResult filters out certain Nodes, the scheduling framework now categorizes them as rejected via `UnschedulableAndUnresolvable`. Consequently, these nodes are excluded from the candidates for the preemption process. Additionally, this update corrects how the scheduling framework handles the Unschedulable status from PreFilter. Previously, if PreFilter returned `Unschedulable`, it could lead to an unexpected abortion in the preemption...

- kubelet: fixed an issue mounting CSI volumes on Windows nodes in 1.32.0 release candidates. ([#129083](https://github.com/kubernetes/kubernetes/pull/129083) [liggitt](https://github.com/liggitt)) [SIG API Machinery, architecture, auth, cli, cloud-provider, cluster-lifecycle, instrumentation,network,node, release, storage, windows ]

- External facing API podresources is now available under k8s.io/kubelet/pkg/apis/ ([#92632](https://github.com/kubernetes/kubernetes/pull/92632), [@RenaudWasTaken](https://github.com/RenaudWasTaken)) [SIG Node and Testing]
- Fewer candidates are enumerated for preemption to improve performance in large clusters. ([#94814](https://github.com/kubernetes/kubernetes/pull/94814), [@adtac](https://github.com/adtac))

- Fixes a regression in 1.34 release candidates where object storage count metrics could include all instances of all types instead of only the object count for a particular type ([#133604](https://github.com/kubernetes/kubernetes/pull/133604), [@serathius](https://github.com/serathius)) [SIG...

    resampling of images.

## Breaking Changes to the API

*   `tf.RewriterConfig` was removed from the Python API after being available in
    1.2 release candidates (it was never in an actual release). Graph rewriting
    is still available, just not as `tf.RewriterConfig`. Instead add an explicit
    import.
*   Breaking change to `tf.contrib.data.Dataset` APIs that expect a nested