},
		{
			name:        "failed to Reserve on node",
			numAllNodes: 1,
			diagnosis: Diagnosis{
				NodeToStatusMap: NodeToStatusMap{
					// There should be only one node here.
					"node1": NewStatus(Unschedulable, "Node failed Reserve plugin Reserve-1"),
				},
			},
			wantReasonMsg: "0/1 nodes are available: 1 Node failed Reserve plugin Reserve-1.",
		},
	}
	for _, tt := range tests {

  MlirOptimizationPassState overall_state = MlirOptimizationPassState::Disabled;

  // Cache per pass state and reuse it during pass execution.
  std::vector<MlirOptimizationPassState> per_pass_state;
  per_pass_state.reserve(registry_->passes().size());

  int num_passes_enabled = 0, num_passes_disabled = 0,
      num_passes_fallback_enabled = 0;
  for (const auto& pass_registration : registry_->passes()) {

		return ScheduleResult{nominatingInfo: clearNominatedNode}, assumedPodInfo, framework.AsStatus(err)
	}

	// Run the Reserve method of reserve plugins.
	if sts := fwk.RunReservePluginsReserve(ctx, state, assumedPod, scheduleResult.SuggestedHost); !sts.IsSuccess() {
		// trigger un-reserve to clean up state associated with the reserved Pod
		fwk.RunReservePluginsUnreserve(ctx, state, assumedPod, scheduleResult.SuggestedHost)

      // Convert values.
      std::vector<Eigen::half> new_values;
      const DenseFPElementsAttr value_attr =
          mlir::cast<DenseFPElementsAttr>(op.getValue());
      new_values.reserve(value_attr.getNumElements());

      for (const float value : value_attr.getValues<float>()) {
        new_values.push_back(Eigen::half(
            std::min(std::max(value, kMinFloat16Value), kMaxFloat16Value)));

            const tensorflow::protobuf::RepeatedPtrField<
                tensorflow::OpDef::ArgDef>& args,
            llvm::StringRef attr_name) {
          std::vector<int32_t> values;
          values.reserve(args.size());
          for (const auto& arg : args) {
            auto range = arg_ranges.at(arg.name());
            values.push_back(range.second - range.first);
          }
          auto attr_value =

            this.serializer = serializer;
        }
    }

    private static class TaggedTypeSerializer<T> extends AbstractSerializer<T> {
        private static final int JAVA_TYPE = 1; // Reserve 0 for null (to be added later)
        private static final TypeInfo JAVA_SERIALIZATION = new TypeInfo(JAVA_TYPE, true, new DefaultSerializer<Object>());

                                         getElementTypeOrSelf(filter_type))
                                         .getScales();
    std::vector<double> result_scales;
    result_scales.reserve(filter_scales.size());

    for (const double filter_scale : filter_scales) {
      result_scales.push_back(input_scale * filter_scale);
    }

    const ArrayRef<int64_t> zero_points =

			}
		}
		if claim.Status.Allocation != nil &&
			!resourceclaim.IsReservedForPod(pod, claim) &&
			resourceclaim.CanBeReserved(claim) {
			// Need to reserve it.
			return true, "need to reserve claim for pod"
		}
	}

	return false, "nothing to do"
}

func (ec *Controller) enqueueResourceClaim(logger klog.Logger, obj interface{}, deleted bool) {

    auto inputs = func.getFunctionType().getInputs();
    entry->addArguments(inputs, SmallVector<Location>(inputs.size(), loc));
    args = entry->getArguments();
  }
  llvm::SmallVector<Value, 4> casted_args;
  casted_args.reserve(func.getNumArguments());
  for (const auto& ArgAndType : zip(args, func.getFunctionType().getInputs())) {
    Value arg = std::get<0>(ArgAndType);
    Type expected_type = std::get<1>(ArgAndType);

                              const tensorflow::TensorShape& shape) {
  static char empty;
  int64_t nelems = 1;
  std::vector<int64_t> dims;
  auto shape_dims = shape.dims();
  dims.reserve(shape_dims);
  for (int i = 0; i < shape_dims; ++i) {
    dims.push_back(shape.dim_size(i));
    nelems *= shape.dim_size(i);
  }
  CHECK_EQ(nelems, 0);
  return TF_NewTensor(