return nullptr;
#else
  tensorflow::ServerDef server_def;
  if (!server_def.ParseFromArray(proto, static_cast<int>(proto_len))) {
    status->status = InvalidArgument(
        "Could not parse provided bytes into a ServerDef protocol buffer");
    return nullptr;
  }

  std::unique_ptr<tensorflow::ServerInterface> out_server;
  status->status = tensorflow::NewServer(server_def, &out_server);

  TF_FilesystemPluginInfo info;
  TF_InitPlugin(&info);
  Status status = filesystem_registration::RegisterFilesystemPluginImpl(&info);
  if (!status.ok()) {
    VLOG(0) << "Static POSIX filesystem could not be registered: " << status;
    return false;
  }
  return true;
}

// Perform the actual registration
static bool unused = StaticallyRegisterLocalFilesystems();

  TF_RETURN_IF_ERROR(
      ops::IdentityN(ctx, inputs, absl::MakeSpan(temp_outputs), "IdentityN"));
  // Although, `ops::IdentityN` returns 2 tensors, the first tensor isn't needed
  // for computing gradient so we could safely drop it.
  outputs[0] = temp_outputs[1];
  temp_outputs[0]->Unref();
  return absl::OkStatus();
}

class CppGradients
    : public ::testing::TestWithParam<std::tuple<const char*, bool, bool>> {
 protected:

  const char* filename;
  int fd;
} PosixFile;

static void Cleanup(TF_RandomAccessFile* file) {
  auto posix_file = static_cast<PosixFile*>(file->plugin_file);
  close(posix_file->fd);
  // This would be safe to free using `free` directly as it is only opaque.
  // However, it is better to be consistent everywhere.
  plugin_memory_free(const_cast<char*>(posix_file->filename));
  delete posix_file;
}

    // case, we will set read to `0` and continue.
    read = 0;
  } else if (!absl::SimpleAtoi(content_length->second, &read)) {
    TF_SetStatus(status, TF_UNKNOWN, "Could not get content-length header");
    return -1;
  }
  // `TF_OUT_OF_RANGE` isn't considered as an error. So we clear it here.
  TF_SetStatus(status, TF_OK, "");

                        const int in_flight_nodes_limit)
      : status_(TF_NewStatus()),
        // If the context's default exector is set to async, re-using that in
        // each thread would cause collectives to deadlock. For consistency we
        // create a new sync executor for every thread.
        //
        // TODO(allenl): We should have an async API that works with the
        // parallel device.

  tensorflow::FunctionDef def;
  CHECK(tensorflow::protobuf::TextFormat::ParseFromString(
      VariableAddFunctionSignature(), &def));
  return def.SerializeAsString();
}

// A graph optimization pass that would fail when triggered for more than once.
class GraphErrorInjectionPass : public tensorflow::GraphOptimizationPass {
 public:
  static bool enabled_;
  GraphErrorInjectionPass() {}

  tensorflow::Status Run(

#include "tensorflow/core/platform/env.h"
#include "tensorflow/core/platform/errors.h"

namespace tensorflow {

// Checks that all schemes provided by a plugin are valid.
// TODO(b/139060984): More validation could be done here, based on supported
// charset, maximum length, etc. Punting it for later.
static Status ValidateScheme(const char* scheme) {
  if (scheme == nullptr)
    return errors::InvalidArgument(

  // `gradient_checker` only works with model that returns only 1 tensor.
  // Although, `ops::SparseSoftmaxCrossEntropyWithLogits` returns 2 tensors, the
  // second tensor isn't needed for computing gradient so we could safely drop
  // it.
  outputs[0] = loss;
  backprop->Unref();
  return absl::OkStatus();
}

Status BiasAddModel(AbstractContext* ctx,
                    absl::Span<AbstractTensorHandle* const> inputs,

      tensorflow::unwrap(h);
  // TODO(b/175427838): It would be nice to be able to use tensorflow::isa here.
  if (tensorflow::CustomDeviceTensorHandle::classof(unwrapped_handle)) {
    return tensorflow::down_cast<tensorflow::CustomDeviceTensorHandle*>(
               unwrapped_handle)
        ->DevicePointer();
  }
  // TODO(b/175427838): It would be nice to be able to use tensorflow::isa here.