auto it =
        shared_name_to_arg_idx.insert({shared_name, block.getNumArguments()});
    if (it.second) {
      auto resource_type = op.getResult(0).getType();
      op.getResult(0).replaceAllUsesWith(
          block.addArgument(resource_type, op.getLoc()));
      AddEntryFunctionInput(
          absl::StrCat("hash_table_", it.first->getValue(), ":0"), target_func);

			arn: ARN{
				Partition:    "minio",
				Service:      "iam",
				Region:       "us-east-1",
				ResourceType: "role",
				ResourceID:   "my-role",
			},
			want: "arn:minio:iam:us-east-1::role/my-role",
		},
		{
			arn: ARN{
				Partition:    "minio",
				Service:      "",
				Region:       "us-east-1",
				ResourceType: "role",
				ResourceID:   "my-role",
			},
			want: "arn:minio::us-east-1::role/my-role",
		},

		Region:       serverRegion,
		ResourceType: arnResourceTypeRole,
		ResourceID:   resourceID,
	}, nil
}

// String - returns string representation of the ARN.
func (arn ARN) String() string {
	return strings.Join(
		[]string{
			arnPrefixArn,
			arn.Partition,
			arn.Service,
			arn.Region,
			"", // account-id is always empty in this implementation
			arn.ResourceType + "/" + arn.ResourceID,
		},
		":",
	)

  } else {
    first_operand_type = op.getOperand().getType();
  }

  auto element_type = getElementTypeOrSelf(first_operand_type);
  if (mlir::isa<TF::ResourceType>(element_type)) {
    first_operand_type =
        mlir::cast<TF::ResourceType>(element_type).getSubtypes().front();
  }

  auto tensor_type = mlir::dyn_cast_or_null<TensorType>(first_operand_type);
  if (!(tensor_type && tensor_type.hasRank())) {

  }

  for (mlir::Value arg : getArgs()) {
    if (mlir::isa<mlir::TF::ResourceType>(
            mlir::getElementTypeOrSelf(arg.getType()))) {
      return emitOpError()
             << "does not support passing '!tf.resource' values as arguments";
    }
  }

  for (mlir::Value result : getResults()) {
    if (mlir::isa<mlir::TF::ResourceType>(
            mlir::getElementTypeOrSelf(result.getType()))) {

  llvm::BitVector removed_params(num_params);
  // Copy the resource-type parameters to the end.
  for (unsigned i = 0; i < num_params; ++i) {
    BlockArgument param = callee.getArgument(i);
    if (mlir::isa<TF::ResourceType>(getElementTypeOrSelf(param.getType()))) {
      removed_params.set(i);
      callee.getBody().addArgument(param.getType(), param.getLoc());
      param.replaceAllUsesWith(callee.getArguments().back());

}

void MarkResourceAsReadAndWrite(
    Value value,
    SmallVectorImpl<SideEffects::EffectInstance<MemoryEffects::Effect>>&
        effects) {
  if (value.getType().cast<TensorType>().getElementType().isa<ResourceType>()) {
    effects.emplace_back(MemoryEffects::Read::get(), value,
                         ResourceEffects::Variable::get());
    effects.emplace_back(MemoryEffects::Write::get(), value,

        def outputDir = file("unpack")
        def sourceFile = file(source)

        buildFile << """
            task unpack(type: Sync) {
                from($type(resources.$resourceType("${sourceFile.toURI()}")))
                into("unpack")
            }
        """

        when:
        packDir.file("A").text = "original"
        packDir."$packType"(sourceFile)

        then:

// limitations under the License.

package config

import (
	"encoding/json"
	"errors"
	"fmt"
	"math"
	"strings"
	"time"
)

type ResourceType int

const (
	Namespace ResourceType = iota
	Deployment
	Pod
	Label
	Annotation
	Container
)

// SelectionSpec is a spec for pods that will be Include in the capture
// archive. The format is:
//

    // unwrap until we get to the inner tensor<...>.
    auto tensor =
        llvm::dyn_cast<TensorType>(var_handle_op.getResource().getType());
    if (!tensor) return nullptr;
    TF::ResourceType resource =
        llvm::dyn_cast<TF::ResourceType>(tensor.getElementType());
    if (!resource || resource.getSubtypes().size() != 1) return nullptr;
    return resource.getSubtypes().front();
  } else if (auto global_tensor_op =