}

TEST_F(CApiExperimentalFunctionTest, EmptyGraphRemoveNonExistentFunction) {
  TF_GraphRemoveFunction(func_graph_, "wrong_name", s_);
  EXPECT_EQ(TF_INVALID_ARGUMENT, TF_GetCode(s_));
  EXPECT_EQ(string("Tried to remove non-existent function 'wrong_name'."),
            string(TF_Message(s_)));
}

TEST_F(CApiExperimentalFunctionTest, GraphRemoveNonExistentFunction) {
  TF_Function* funcs[1];
  DefineFunction(func_name_, &func_);

    // Returning null tensors from a tf.function is not supported, so we keep
    // track of indices in the model's outputs are nullptr in this set.
    // The FunctionDef only outputs the non-null tensors. We later pad the
    // function op outputs to have nullptrs at the `null_indices`.
    absl::flat_hash_set<int> null_indices;
    {
      AbstractContextPtr func_ctx(BuildFunction(fn_name));

                      &FloatDeallocator, nullptr);
}

// All the *Helper methods are used as a workaround for the restrictions that
// one cannot call ASSERT_* methods in non-void-returning functions (when
// exceptions are disabled during compilation)
void PlaceholderHelper(TF_Graph* graph, TF_Status* s, const char* name,
                       TF_DataType dtype, const std::vector<int64_t>& dims,

#   value in the "env_vars" list that you can choose to copy that environment.
#      Ex. 1: TFCI=py311,linux_x86_cuda,nightly_upload  (nightly job)
#      Ex. 2: TFCI=py39,linux_x86,rbe                   (continuous job)
#   Non-Googlers should replace "nightly_upload" or "rbe" with
#   "public_cache,disk_cache".
#   Googlers should replace "nightly_upload" with "public_cache,disk_cache" or

// one of the source or destination devices to be CPU (i.e., for the source or
// destination tensor to be placed in host memory).
// If async execution is enabled, the copy may be enqueued and the call will
// return "non-ready" handle. Else, this function returns after the copy has
// been done.
TF_CAPI_EXPORT extern TFE_TensorHandle* TFE_TensorHandleCopyToDevice(
    TFE_TensorHandle* h, TFE_Context* ctx, const char* device_name,

  # By default, aliases are only expanded in interactive shells, which means
  # that they are not substituted for their corresponding commands in shell
  # scripts. By setting "expand_aliases", we enable alias expansion in
  # non-interactive shells as well.
  shopt -s expand_aliases
else
  echo '==TFCI==: Error: Cannot find path to grealpath or gstat'
  echo 'TF CI scripts require GNU core utilties to be installed. Please make'

                  TF_Status* s, const char* name = "min");

TF_Operation* Mul(TF_Operation* l, TF_Operation* r, TF_Graph* graph,
                  TF_Status* s, const char* name = "mul");

// If `op_device` is non-empty, set the created op on that device.
TF_Operation* MinWithDevice(TF_Operation* l, TF_Operation* r, TF_Graph* graph,
                            const string& op_device, TF_Status* s,

// Monitoring Counter APIs.
// These APIs de-templated monitoring Counter for swig.

typedef struct TFE_MonitoringCounterCell TFE_MonitoringCounterCell;

// Atomically increments the value of the cell. The value must be non-negative.
TF_CAPI_EXPORT extern void TFE_MonitoringCounterCellIncrementBy(
    TFE_MonitoringCounterCell* cell, int64_t value);

// Retrieves the current value of the cell.

      dynamic_addr(symbol) = static_addr(symbol) + load_addr
 
-     The choice of symbol is arbitrary. The static address we obtain
-     by constructing a non GOT reference to the symbol, the dynamic
-     address of the symbol we compute using adrp/add to compute the
-     symbol's address relative to the PC. */
-
-  ElfW(Addr) static_addr;
-  ElfW(Addr) dynamic_addr;
-

*   The `tf.keras.mixed_precision.experimental` API has been removed. The non-experimental symbols under `tf.keras.mixed_precision` have been available since TensorFlow 2.4 and should be used instead.
    * The non-experimental API has some minor differences from the experimental API. In most cases, you only need to make three minor changes: