%1:5 = "tf.PartitionedCall"(%arg0, %arg1, %arg2, %arg3, %arg4, %arg5) {Tin = ["tfdtype$DT_FLOAT", "tfdtype$DT_FLOAT", "tfdtype$DT_FLOAT", "tfdtype$DT_FLOAT", "tfdtype$DT_FLOAT", "tfdtype$DT_FLOAT"], Tout = ["tfdtype$DT_FLOAT", "tfdtype$DT_FLOAT", "tfdtype$DT_FLOAT", "tfdtype$DT_FLOAT", "tfdtype$DT_FLOAT"], _output_shapes = ["tfshape$dim { size: 9 } dim {...

func.func @packStringWithFlex(%arg0: tensor<2x!tf_type.string>, %arg1: tensor<2x!tf_type.string>) -> tensor<2x2x!tf_type.string> {
  %0 = "tf.Pack"(%arg0, %arg1) : (tensor<2x!tf_type.string>, tensor<2x!tf_type.string>) -> tensor<2x2x!tf_type.string>
  func.return %0 : tensor<2x2x!tf_type.string>

// CHECK-LABEL: packStringWithFlex
// CHECK: "tf.Pack"(%arg0, %arg1) : (tensor<2x!tf_type.string>, tensor<2x!tf_type.string>) -> tensor<2x2x!tf_type.string>
}

func.func @testNoAddToAddV2ForStringType(%arg0: tensor<8x16x!tf_type.string>, %arg1: tensor<8x16x!tf_type.string>) -> tensor<8x16x!tf_type.string> {
  %0 = "tf.Add"(%arg0, %arg1) : (tensor<8x16x!tf_type.string>, tensor<8x16x!tf_type.string>) -> tensor<8x16x!tf_type.string>
  func.return %0: tensor<8x16x!tf_type.string>

  // CHECK: %0 = "tf.Add"(%arg0, %arg1) : (tensor<8x16x!tf_type.string>, tensor<8x16x!tf_type.string>) -> tensor<8x16x!tf_type.string>

    ```mlir
    func.func @single_state_single_dataset_type_no_arguments(%arg0: tensor<!tf_type.variant>, %arg1: tensor<i64>) {
      %0 = "tf.AnonymousIteratorV3"() {output_shapes = [#tf_type.shape<32>], output_types = [f32]} : () -> tensor<!tf_type.resource>
      "tf.MakeIterator"(%arg0, %0) : (tensor<!tf_type.variant>, tensor<!tf_type.resource>) -> ()
      %cst = "tf.Const"() {value = dense<true> : tensor<i1>} : () -> tensor<i1>

    // CHECK:      metadata
    // CHECK:      dtype: DT_RESOURCE
    // CHECK-SAME: kind: VARIABLE

    func.return %0: tensor<*x!tf_type.resource>
  }
  func.func @_func(%arg0: tensor<*x!tf_type.resource>) -> tensor<*x!tf_type.resource> {
    func.return %arg0 : tensor<*x!tf_type.resource>
  }
}

// -----

func.func @testGatherUnsupported2DStringInput(%arg0 : tensor<?x?x!tf_type.string>, %arg1 : tensor<?xi32>) -> tensor<*x!tf_type.string> {
  // expected-error @+1 {{'tfl.gather' op expect 1d input when the given type is string, got 'tensor<?x?x!tf_type.string>'}}
  %0 = "tfl.gather"(%arg0, %arg1) {axis = 1 : i32}: (tensor<?x?x!tf_type.string>,tensor<?xi32>) -> tensor<*x!tf_type.string>
  func.return %0 : tensor<*x!tf_type.string>
}

// -----

// This method will actually merge the information contained in the
// types, it is capable of refining:
//   tensor<!tf_type.variant<tensor<?x8xf32>>>
// and:
//   tensor<!tf_type.variant<tensor<10x?xf32>>>
// into:
//   tensor<!tf_type.variant<tensor<10x8xf32>>>
//
// In case of inconsistencies (rank disagreement for example), it returns `lhs`.
Type TypeMeet(Type lhs, Type rhs) {

	handle = Handle(r0)
	if handle == 0 {
		err = errnoErr(e1)
	}
	return
}

func createWellKnownSid(sidType WELL_KNOWN_SID_TYPE, domainSid *SID, sid *SID, sizeSid *uint32) (err error) {
	r1, _, e1 := syscall.Syscall6(procCreateWellKnownSid.Addr(), 4, uintptr(sidType), uintptr(unsafe.Pointer(domainSid)), uintptr(unsafe.Pointer(sid)), uintptr(unsafe.Pointer(sizeSid)), 0, 0)
	if r1 == 0 {
		err = errnoErr(e1)
	}

        }
        node->AddAttr("shape", index_it->second->second.shape);
        DataType dtype = index_it->second->second.imported_dtype;
        // Uses the existing output type if it isn't specified by the user.
        if (dtype == DT_INVALID) {
          dtype = node->output_type(0);
        }
        node->AddAttr("dtype", dtype);
      }
    }
    if (!node_added_to_shape_refiner) {

    return tflite::TensorType_UINT8;
  } else if (auto complex_type = mlir::dyn_cast<mlir::ComplexType>(type)) {
    auto ftype = complex_type.getElementType();
    if (ftype.isF32()) {
      return tflite::TensorType_COMPLEX64;
    }
    if (ftype.isF64()) {
      return tflite::TensorType_COMPLEX128;
    }
    return Status(absl::StatusCode::kInvalidArgument, "Unsupported type");