- [Container Images](#container-images-1)
  - [Changelog since v1.21.12](#changelog-since-v12112)
  - [Changes by Kind](#changes-by-kind)
    - [Bug or Regression](#bug-or-regression)
  - [Dependencies](#dependencies-1)
    - [Added](#added-1)
    - [Changed](#changed-1)
    - [Removed](#removed-1)
- [v1.21.12](#v12112)
  - [Downloads for v1.21.12](#downloads-for-v12112)
    - [Source Code](#source-code-2)

      output_batch_dimension = 0,
      output_feature_dimension = 3,
      output_spatial_dimensions = [1, 2]
    >, feature_group_count = 1 : i64, lhs_dilation = dense<2> : tensor<2xi64>, padding = dense<[[2, 1], [2, 1]]> : tensor<2x2xi64>,
    precision_config = [#mhlo<precision DEFAULT>, #mhlo<precision DEFAULT>], rhs_dilation = dense<1> : tensor<2xi64>, window_strides = dense<1> : tensor<2xi64>

	/*2121*/ uint16(xReadSlashR),
	/*2122*/ uint16(xArgXmm1),
	/*2123*/ uint16(xArgXmm2M32),
	/*2124*/ uint16(xMatch),
	/*2125*/ uint16(xSetOp), uint16(UCOMISD),
	/*2127*/ uint16(xReadSlashR),
	/*2128*/ uint16(xArgXmm1),
	/*2129*/ uint16(xArgXmm2M64),
	/*2130*/ uint16(xMatch),
	/*2131*/ uint16(xCondPrefix), 2,
	0x66, 2143,
	0x0, 2137,
	/*2137*/ uint16(xSetOp), uint16(COMISS),
	/*2139*/ uint16(xReadSlashR),

         <match value="\n@article{" type="stringignorecase" offset="2:128"/>
         <match value="\n@book{" type="stringignorecase" offset="2:128"/>
         <match value="\n@inbook{" type="stringignorecase" offset="2:128"/>
         <match value="\n@incollection{" type="stringignorecase" offset="2:128"/>
         <match value="\n@inproceedings{" type="stringignorecase" offset="2:128"/>

  // CHECK: "mhlo.reduce_window"(%[[ARG]], %[[INIT]])
  // CHECK: <{window_dimensions = dense<[1, 2, 2, 1]> : tensor<4xi64>, window_strides = dense<[1, 4, 4, 1]> : tensor<4xi64>}>
  // CHECK: mhlo.maximum
  // CHECK: mhlo.return

  %0 = "tf.MaxPool"(%arg0) {data_format = "NHWC", ksize = [1, 2, 2, 1], padding = "VALID", strides = [1, 4, 4, 1]} : (tensor<2x12x20x7xi32>) -> tensor<2x3x5x7xi32>

  %211:2 = "tf.RaggedRange"(%210, %207, %13) {T = i64, Tsplits = i64, device = ""} : (tensor<?xi64>, tensor<?xi64>, tensor<i64>) -> (tensor<?xi64>, tensor<?xi64>)

\"x-prop\": {\"rule\": \"self.x__dash__prop > 0\"}\n  - Rule accessing a property named \"redact__d\": {\"rule\": \"self.redact__underscores__d > 0\"}\n\nEquality on arrays with x-kubernetes-list-type of 'set' or 'map' ignores element order, i.e. [1, 2] == [2, 1]. Concatenation on arrays with x-kubernetes-list-type use the semantics of the list type:\n  - 'set': `X + Y` performs a union where the array positions of all elements in `X` are preserved and\n    non-intersecting elements in `Y` are appended,...

* ♻ Strip empty whitespace from description extracted from docstrings. PR [#2821](https://github.com/tiangolo/fastapi/pull/2821) by [@and-semakin](https://github.com/and-semakin).

kubernetes/enhancements: [#121](https://github.com/kubernetes/enhancements/issues/121#issuecomment-457396290) [[kep](https://github.com/kubernetes/enhancements/blob/master/keps/sig-storage/121-local-persistent-volumes/README.md)]

Pid Limiting is Graduating to Beta ([#757](https://github.com/kubernetes/enhancements/issues/757))

	ap_ImmUnsigned_16_31_48_63       = &argField{Type: TypeImmUnsigned, Shift: 0, BitFields: BitFields{{16, 16, 0}, {16, 16, 1}}}
	ap_ImmUnsigned_21_22             = &argField{Type: TypeImmUnsigned, Shift: 0, BitFields: BitFields{{21, 2, 0}}}
	ap_ImmUnsigned_18_20             = &argField{Type: TypeImmUnsigned, Shift: 0, BitFields: BitFields{{18, 3, 0}}}
	ap_ImmUnsigned_19_20             = &argField{Type: TypeImmUnsigned, Shift: 0, BitFields: BitFields{{19, 2, 0}}}