* is a power of two. Indexing uses masked per-thread hash codes.
   * Nearly all declarations in this class are package-private,
   * accessed directly by subclasses.
   *
   * Table entries are of class Cell; a variant of AtomicLong padded
   * to reduce cache contention on most processors. Padding is
   * overkill for most Atomics because they are usually irregularly
   * scattered in memory and thus don't interfere much with each

    # Dev builds from prow are on gcr.io
    hub: gcr.io/istio-testing
    # Default tag for Istio images.
    tag: latest
    # Variant of the image to use.
    # Currently supported are: [debug, distroless]
    variant: ""

    # Specify image pull policy if default behavior isn't desired.
    # Default behavior: latest images will be Always else IfNotPresent.
    imagePullPolicy: ""

    # Dev builds from prow are on gcr.io
    hub: gcr.io/istio-testing
    # Default tag for Istio images.
    tag: latest
    # Variant of the image to use.
    # Currently supported are: [debug, distroless]
    variant: ""
    # Specify image pull policy if default behavior isn't desired.
    # Default behavior: latest images will be Always else IfNotPresent.
    imagePullPolicy: ""

  "$python" -c 'import sys; import tensorflow as tf; sys.exit(0 if "keras" in tf.keras.__name__ else 1)'
fi
# VERY basic check to ensure the [and-cuda] package variant is installable.
# Checks TFCI_BAZEL_COMMON_ARGS for "gpu" or "cuda", implying that the test is
# relevant. All of the GPU test machines have CUDA installed via other means,

   * is a power of two. Indexing uses masked per-thread hash codes.
   * Nearly all declarations in this class are package-private,
   * accessed directly by subclasses.
   *
   * Table entries are of class Cell; a variant of AtomicLong padded
   * to reduce cache contention on most processors. Padding is
   * overkill for most Atomics because they are usually irregularly
   * scattered in memory and thus don't interfere much with each

      - name: istio-proxy
{{- if contains "/" .Values.image }}
        image: "{{ .Values.image }}"
{{- else }}
        image: "{{ .Values.hub }}/{{ .Values.image | default "ztunnel" }}:{{ .Values.tag }}{{with (.Values.variant )}}-{{.}}{{end}}"
{{- end }}
        ports:
        - containerPort: 15020
          name: ztunnel-stats
          protocol: TCP
        resources:
{{- if .Values.resources }}

    # Dev builds from prow are on gcr.io
    hub: gcr.io/istio-testing

    # Default tag for Istio images.
    tag: latest

    # Variant of the image to use.
    # Currently supported are: [debug, distroless]
    variant: ""

    # Specify image pull policy if default behavior isn't desired.
    # Default behavior: latest images will be Always else IfNotPresent.
    imagePullPolicy: ""

    return tensorflow::ops::Identity(ctx, input, output, name);
  } else if (!DataTypeIsComplex(BaseType(dtype)) &&
             BaseType(dtype) != DT_VARIANT) {
    return errors::InvalidArgument(
        "Expected numeric or variant tensor, got dtype ", dtype);
  }
  return tensorflow::ops::Conj(ctx, input, output, name);
}

class AddGradientFunction : public GradientFunction {
 public:

`kube.Client` just bundles these all in one place, so we have a single object to pass around anywhere we need Kubernetes access.
Additionally, it has a fake client variant, that does the same thing but with fake clients for use in unit tests.

Aside from this, there are a few conveniences and workarounds built-in to the client to make things simpler.

        if (executer == defaultExecuter) {
            // The test task with the default executer runs with 'check'
            tasks.named("check").configure { dependsOn(testTask) }
        }
    }
    // Create a variant of the test suite to force realization of component metadata
    if (testType == TestType.INTEGRATION) {
        createTestTask(prefix + "ForceRealizeTest", defaultExecuter, sourceSet, testType) {