TFE_DeleteContext(ctx);
  CHECK_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);
  TF_DeleteStatus(status);
}
BENCHMARK(BM_InitOp);

void BM_Execute(::testing::benchmark::State& state) {
  const int async = state.range(0);
  state.SetLabel(async ? "ExecuteAsync" : "Execute");
  TF_Status* status = TF_NewStatus();
  TFE_ContextOptions* opts = TFE_NewContextOptions();
  TFE_ContextOptionsSetAsync(opts, static_cast<unsigned char>(async));

            static_cast<TF_Code>(tensorflow::error::FAILED_PRECONDITION));
  EXPECT_EQ(TF_ABORTED, static_cast<TF_Code>(tensorflow::error::ABORTED));
  EXPECT_EQ(TF_OUT_OF_RANGE,
            static_cast<TF_Code>(tensorflow::error::OUT_OF_RANGE));
  EXPECT_EQ(TF_UNIMPLEMENTED,
            static_cast<TF_Code>(tensorflow::error::UNIMPLEMENTED));
  EXPECT_EQ(TF_INTERNAL, static_cast<TF_Code>(tensorflow::error::INTERNAL));

  StringPiece result;
  // read at least 1 byte more than test_data
  status = read_file->Read(0, test_data.size() + 1, &result, scratch);
  EXPECT_PRED2(UnimplementedOrReturnsCode, status, Code::OUT_OF_RANGE);
}

// The URI schemes that need to be tested are provided by the user via flags
// (or, if none is supplied, all existing schemes are used). As a scheme can

        session->graph->mu.unlock();
        return false;
      }

      GraphDef graph_def;
      *graph_def.mutable_versions() = graph.versions();
      // Fill graph_def with nodes with ids in the range
      // [session->last_num_graph_nodes, num_nodes), that is the nodes
      // added since the last TF_SessionRun() call.
      for (auto id = session->last_num_graph_nodes; id < num_nodes; ++id) {

          &(gcs_file->buffer[0]), status);
      gcs_file->buffer_end_is_past_eof =
          (TF_GetCode(status) == TF_OUT_OF_RANGE);
      if (read_fill_buffer >= 0) gcs_file->buffer.resize(read_fill_buffer);
      if (TF_GetCode(status) != TF_OK &&
          TF_GetCode(status) != TF_OUT_OF_RANGE) {
        // Empty the buffer to avoid caching bad reads.
        gcs_file->buffer.resize(0);
        return -1;
      }

to the network can execute arbitrary code with the privileges of the user
running the `tf.train.Server`.

## Untrusted inputs during training and prediction

TensorFlow supports a wide range of input data formats. For example it can
process images, audio, videos, and text. There are several modules specialized
in taking those formats, modifying them, and/or converting them to intermediate

/// don't need to read the `.cc` file but they should consult every section of
/// this file to ensure a compliant plugin can be built and that the plugin can
/// be used without recompilation in the widest range of TensorFlow versions.
///
/// The header is divided into sections, as follows:
///   1. Opaque plugin private data structures and wrappers for type safety;
///   2. Function tables for plugin functionality;

        `DynamicUpdateSlice`.
    *   Enabled a new MLIR-based dynamic range quantization backend by default
        *   The new backend is used for post-training int8 dynamic range
            quantization and post-training float16 quantization.
        *   Set `experimental_new_dynamic_range_quantizer` in
            tf.lite.TFLiteConverter to False to disable this change

  TF_Operation* feed2 = Placeholder(func_graph_, s_, "feed2");
  TF_Operation* add = Add(feed1, feed2, func_graph_, s_);
  DefineT(-1, {}, {{feed1, 0}, {feed2, 2}}, {{add, 0}}, {}, true);
  EXPECT_EQ(TF_OUT_OF_RANGE, TF_GetCode(s_));
  EXPECT_EQ(string("Node 'feed2' (type: 'Placeholder', num of outputs: 1) does "
                   "not have output 2\n\tEncountered while processing "
                   "input 1 into function 'MyFunc'"),

          echo "TF_VERSION=${major_version}.${minor_version}" >> "$GITHUB_OUTPUT"
          # Also get the current date to do cache busting. Assumes one day
          # is an ok range for rebuilds
          echo "DATE=$(date +'%Y-%m-%d')" >> "$GITHUB_OUTPUT"
        id: tf-version
      -
        name: Build and push
        id: docker_build