}
  }
  attr {
    key: "shape"
    value {
      shape {
      }
    }
  }
  attr {
    key: "shared_name"
    value {
      s: ""
    }
  }
}
node {
  name: "variable_v2_node"
  op: "VariableV2"
  attr {
    key: "dtype"
    value {
      type: DT_INT64
    }
  }
  attr {
    key: "shape"
    value {
      shape {
      }
    }
  }
  attr {

node: {
  name: "x"
  op: "Placeholder"
  attr: {
    key: "shape"
    value: {
      shape: {
        dim: {
          size: -1
        }
      }
    }
  }
  attr: {
    key: "dtype"
    value: {
      type: DT_INT32
    }
  }
}
node: {
  name: "y"
  op: "Placeholder"
  attr: {
    key: "shape"
    value: {
      shape: {
        dim: {
          size: -1
        }
      }

// CHECK-NEXT:   subgraphs: [ {
// CHECK-NEXT:     tensors: [ {
// CHECK-NEXT:       shape: [  ],
// CHECK-NEXT:       type: INT32,
// CHECK-NEXT:       buffer: 1,
// CHECK-NEXT:       name: "arg0",
// CHECK-NEXT:       quantization: {
// CHECK-EMPTY:
// CHECK-NEXT:       },
// CHECK-NEXT:       has_rank: true
// CHECK-NEXT:     }, {
// CHECK-NEXT:       shape: [ 1 ],
// CHECK-NEXT:       buffer: 2,
// CHECK-NEXT:       name: "arg1",

namespace tensorflow {

std::string ImmediateExecutionTensorHandle::DebugString() const {
  PartialTensorShape shape;
  std::string shape_string;
  if (Shape(&shape).ok()) {
    shape_string = shape.DebugString();
  } else {
    shape_string = "<error computing shape>";
  }
  std::string value_string;
  if (!SummarizeValue(value_string).ok()) {
    value_string = "<error computing value>";
  }

    %3 = "tf.RangeDataset"(%0, %1, %2) {device = "/device:CPU:0", output_shapes = [#tf_type.shape<>], output_types = [i64], metadata = ""} : (tensor<i64>, tensor<i64>, tensor<i64>) -> tensor<!tf_type.variant>
    // CHECK: tfrt_fallback_async.executeop key({{[0-9]+}}) cost({{.*}}) device("/device:CPU:0") "tf.FlatMapDataset"({{.*}}) {Targuments = [], metadata = "", output_shapes = [#corert.shape<>], output_types = [i64]} {f = "funcB"} : 1

# Text form of tensorflow.tf2xla.Config proto.
feed {
  id { node_name: "p_hold" }
  shape {}
}
feed {
  id { node_name: "x_hold" }
  shape {
    dim { size: 1 }
  }
}
feed {
  id { node_name: "y_hold" }
  shape {
    dim { size: 1 }
  }
}
fetch {
  id { node_name: "result" }

  TF_ASSERT_OK_AND_ASSIGN(auto stream, executor->CreateStream());

  se::DeviceMemoryBase gpu_dst{device_tensor.data(), 4 * sizeof(float)};
  xla::Shape shape;
  TF_ASSERT_OK(TensorShapeToXLAShape(DT_FLOAT, TensorShape({2, 2}), &shape));

  // Copy the cpu_tensor to the GPU first before trying to copy it back.
  TF_ASSERT_OK(
      stream->Memcpy(&gpu_dst, origin_cpu_tensor.data(), gpu_dst.size()));

  // CHECK-NOT: tf.Assign

  // CHECK: [[handle:%.*]] = "tf.VarHandleOp"
  // CHECK-SAME: shared_name = "x"
  // CHECK: "tf.AssignVariableOp"([[handle]], {{%.*}})
  %0 = "tf.VariableV2"() {container = "", shape = #tf_type.shape<>, shared_name = "x"} : () -> tensor<!tf_type.int32ref>
  %1 = "tf.Const"() {value = dense<0> : tensor<i32>} : () -> tensor<i32>

// CHECK-NEXT:   subgraphs: [ {
// CHECK-NEXT:     tensors: [ {
// CHECK-NEXT:       shape: [ 1 ],
// CHECK-NEXT:       buffer: 1,
// CHECK-NEXT:       name: "arg0",
// CHECK-NEXT:       quantization: {
// CHECK-EMPTY:
// CHECK-NEXT:       },
// CHECK-NEXT:       has_rank: true
// CHECK-NEXT:     }, {
// CHECK-NEXT:       shape: [ 1 ],
// CHECK-NEXT:       buffer: 2,
// CHECK-NEXT:       name: "arg1",

TF_CAPI_EXPORT extern int TF_ShapeDims(const TF_Shape* shape);

// Returns the `d`th dimension of `shape`. If `shape` has unknown rank,
// invoking this function is undefined behavior. Returns -1 if dimension is
// unknown.
TF_CAPI_EXPORT extern int64_t TF_ShapeDimSize(const TF_Shape* shape, int d);

// Deletes `shape`.
TF_CAPI_EXPORT extern void TF_DeleteShape(TF_Shape* shape);

#ifdef __cplusplus
} /* end extern "C" */
#endif