* The main mime type should be the canonical one, use aliases for any
     other widely used forms
   * Where there's a hierarchy in the types, list it via a parent
   * Highly specific magic matches get a high priority
   * General magic matches which could trigger a false-positive need
     a low one
   * The priority for containers normally need to be higher than for
     the things they contain, so they don't accidently get detected

  The scheduling hints allow the scheduler to only retry scheduling a `Pod`
  that was previously rejected by the `NodeAffinity` plugin if a new `Node` or a `Node` update matches the `Pod`'s node affinity. ([#119155](https://github.com/kubernetes/kubernetes/pull/119155), [@carlory](https://github.com/carlory))
- `kubeadm`: promoted feature gate `EtcdLearnerMode` to `beta`. Learner mode for

        "description": "A label selector is a label query over a set of resources. The result of matchLabels and matchExpressions are ANDed. An empty label selector matches all objects. A null label selector matches no objects.",
        "properties": {
          "matchExpressions": {
            "description": "matchExpressions is a list of label selector requirements. The requirements are ANDed.",

### Bug or Regression

- EndpointSlice IP validation now matches Endpoints IP validation. ([#101084](https://github.com/kubernetes/kubernetes/pull/101084), [@robscott](https://github.com/robscott)) [SIG Apps and Network]

- Fixes issue with websocket-based watches of Service objects not closing correctly on timeout ([#102541](https://github.com/kubernetes/kubernetes/pull/102541), [@liggitt](https://github.com/liggitt)) [SIG API Machinery and Testing]

## Dependencies

### Added

* Active watches of custom resources now terminate properly if the CRD is modified. ([#78029](https://github.com/kubernetes/kubernetes/pull/78029), [@liggitt](https://github.com/liggitt))

* Fix incorrect conflict errors applying strategic merge patches to resources. ([#43871](https://github.com/kubernetes/kubernetes/pull/43871), [@liggitt](https://github.com/liggitt))

    // The activations (inputs) form the LHS of the convolution.
    // Activations have shape: [batch, in_rows, in_cols, ..., in_depth]
    // For the gradient computation, we need to:
    // 1. In the case of group convolution, move the num_groups dimension before
    // the batch dimension
    // 2. Swap the roles of the batch and feature dimensions.
    const int feature_dim =

- Kube-apiserver: ensures the namespace of objects sent to admission webhooks matches the request namespace. Previously, objects without a namespace set would have the request namespace populated after mutating admission, and objects with a namespace that did not match the request namespace would be rejected after admission. ([#94...

- Fix: initial delay in mounting azure disk & file ([#93052](https://github.com/kubernetes/kubernetes/pull/93052), [@andyzhangx](https://github.com/andyzhangx)) [SIG Cloud Provider and Storage]
- Fixed a performance issue applying json patches to deeply nested objects ([#93812](https://github.com/kubernetes/kubernetes/pull/93812), [@liggitt](https://github.com/liggitt)) [SIG API Machinery, CLI, Cloud Provider, Cluster Lifecycle and Instrumentation]