// Registers Graph optimizers.
Status InitGraphPlugin(void* dso_handle);

// Allow registering a graph optimizer using a function (used for
// testing).
Status InitGraphPlugin(TFInitGraphPluginFn init_fn);

struct GrapplerItem;
class Cluster;

struct TFStatusDeleter {
  void operator()(TF_Status* s) const { TF_DeleteStatus(s); }
};
using OwnedTFStatus = std::unique_ptr<TF_Status, TFStatusDeleter>;

// or removed during the graph transformation. This includes feed and fetch
// nodes, keep_ops, init_ops. Fills in `num_values` and `storage_size`, they
// will be used in `TF_GetNodesToPreserveList`.
TF_CAPI_EXPORT extern void TF_GetNodesToPreserveListSize(
    const TF_GrapplerItem* item, int* num_values, size_t* storage_size,
    TF_Status* status);

    TFE_Context* ctx, int keep_alive_secs, const void* proto, size_t proto_len,
    int64_t init_timeout_in_ms, TF_Status* status);

// This API is for experimental usage and may be subject to change.
TF_CAPI_EXPORT extern void TFE_ContextSetServerDefWithTimeout(
    TFE_Context* ctx, int keep_alive_secs, const void* proto, size_t proto_len,
    int64_t init_timeout_in_ms, TF_Status* status,

                           bool isolate_session_state,
                           int64_t init_timeout_in_ms);

tensorflow::ServerDef ReplaceTaskInServerDef(
    const tensorflow::ServerDef& server_def, int task_index);

void ReplaceTaskInServerDef(tensorflow::ServerDef* server_def, int task_index,
                            const std::string& host, int port);

std::vector<std::string> ListDeviceNames(TFE_Context* ctx);

#define WIN32_LEAN_AND_MEAN
#include <windows.h>
#include <process.h>
#include "_cgo_export.h"

__stdcall
static unsigned int
addThread(void *p)
{
	int i, max;
	
	max = *(int*)p;
	for(i=0; i<max; i++)
		Add(i);
	return 0;
}

void
doAdd(int max, int nthread)
{
	enum { MaxThread = 20 };
	int i;
	uintptr_t thread_id[MaxThread];
	
	if(nthread > MaxThread)
		nthread = MaxThread;
	for(i=0; i<nthread; i++)

// is needed a static_cast can be applied.
class ImmediateExecutionTensorHandle : public AbstractTensorHandle {
 public:
  // Returns number of dimensions.
  virtual Status NumDims(int* num_dims) const = 0;
  // Returns number of elements across all dimensions.
  virtual Status NumElements(int64_t* num_elements) const = 0;
  // Returns size of specified dimension
  //

void TF_GraphSetOutputHandleShapesAndTypes(TF_Graph* graph, TF_Output output,
                                           int num_shapes_and_types,
                                           const int64_t** shapes,
                                           const int* ranks,
                                           const TF_DataType* types,
                                           TF_Status* status);

// Return a tensor handle with given type, values and dimensions.
template <class T, TF_DataType datatype>
Status TestTensorHandleWithDims(AbstractContext* ctx, const T* data,
                                const int64_t* dims, int num_dims,
                                AbstractTensorHandle** tensor) {
  std::unique_ptr<TF_Status, decltype(&TF_DeleteStatus)> status(
      TF_NewStatus(), TF_DeleteStatus);
  TFE_Context* eager_ctx =

// Prepare tracing to the expected number of output for an operation.
void TF_OutputListSetNumOutputs(TF_OutputList* o, int num_outputs, TF_Status*);
// Return the number of outputs in the list.
int TF_OutputListNumOutputs(TF_OutputList* o);
// Return the `i`th output in the list.
TF_AbstractTensor* TF_OutputListGet(TF_OutputList* o, int i);
// Append a tensor at the end of the output list, growing its size by one.

  virtual bool HasCustomDeviceInput() const = 0;

  virtual const tensorflow::OpDef* OpDef() const = 0;

  virtual Status InputLength(const char* input_name, int* length) = 0;
  virtual Status OutputLength(const char* output_name, int* length) = 0;

  // Set stack trace to be used for potential async error reporting.
  virtual void SetStackTrace(ManagedStackTrace stack_trace) = 0;