[BE] Enforce missing override keyword (pytorch#104032)

This PR enables `-Winconsistent-missing-destructor-override` and `-Winconsistent-missing-override`
and fixes violations.

<!--
copilot:summary
-->
### <samp>🤖 Generated by Copilot at 47e904e</samp>

This pull request updates the code of various classes and operators in the `caffe2` and `aten` subdirectories to use the `override` specifier instead of the `virtual` keyword for destructors and other virtual functions that override a base class function. This improves the code readability, quality, and consistency with C++ best practices. It also modifies the `./CMakeLists.txt` file to enable warnings for these specifiers, but disable errors.

Pull Request resolved: pytorch#104032
Approved by: https://github.com/malfet

CMakeLists.txt

@@ -817,6 +817,8 @@ if(NOT MSVC)
   append_cxx_flag_if_supported("-Wno-strict-aliasing" CMAKE_CXX_FLAGS)
   append_cxx_flag_if_supported("-Wvla-extension" CMAKE_CXX_FLAGS)
   append_cxx_flag_if_supported("-Wnewline-eof" CMAKE_CXX_FLAGS)
+  append_cxx_flag_if_supported("-Winconsistent-missing-override" CMAKE_CXX_FLAGS)
+  append_cxx_flag_if_supported("-Winconsistent-missing-destructor-override" CMAKE_CXX_FLAGS)
   if("${CMAKE_CXX_COMPILER_ID}" MATCHES "Clang")
     string(APPEND CMAKE_CXX_FLAGS " -Wno-range-loop-analysis")
     string(APPEND CMAKE_CXX_FLAGS " -Wno-pass-failed")
@@ -861,6 +863,8 @@ if(NOT MSVC)
 
   append_cxx_flag_if_supported("-Wno-error=pedantic" CMAKE_CXX_FLAGS)
   append_cxx_flag_if_supported("-Wno-error=old-style-cast" CMAKE_CXX_FLAGS)
+  append_cxx_flag_if_supported("-Wno-error=inconsistent-missing-override" CMAKE_CXX_FLAGS)
+  append_cxx_flag_if_supported("-Wno-error=inconsistent-missing-destructor-override" CMAKE_CXX_FLAGS)
   append_cxx_flag_if_supported("-Wconstant-conversion" CMAKE_CXX_FLAGS)
   append_cxx_flag_if_supported("-Wno-invalid-partial-specialization" CMAKE_CXX_FLAGS)
   append_cxx_flag_if_supported("-Wno-unused-private-field" CMAKE_CXX_FLAGS)

aten/src/ATen/test/cpu_rng_test.cpp

@@ -19,8 +19,7 @@ constexpr auto kCustomRNG = DispatchKey::CustomRNGKeyId;
 
 struct TestCPUGenerator : public c10::GeneratorImpl {
   TestCPUGenerator(uint64_t value) : GeneratorImpl{Device(DeviceType::CPU), DispatchKeySet(kCustomRNG)}, value_(value) { }
-  // NOLINTNEXTLINE(modernize-use-override)
-  ~TestCPUGenerator() = default;
+  ~TestCPUGenerator() override = default;
   uint32_t random() { return value_; }
   uint64_t random64() { return value_; }
   c10::optional<float> next_float_normal_sample() { return next_float_normal_sample_; }

caffe2/contrib/gloo/allgather_ops.h

@@ -44,7 +44,7 @@ class AllgatherOp final : public Operator<Context> {
     }
   }
 
-  virtual ~AllgatherOp() {}
+  ~AllgatherOp() override {}
 
   bool RunOnDevice() override {
     std::call_once(once_, [&] { initialize(); });

caffe2/contrib/gloo/allreduce_ops.h

@@ -32,7 +32,7 @@ class AllreduceOp final : public Operator<Context> {
     }
   }
 
-  virtual ~AllreduceOp() {}
+  ~AllreduceOp() override {}
 
   bool RunOnDevice() override {
     std::call_once(once_, [&] { initialize(); });

caffe2/contrib/gloo/barrier_ops.h

@@ -24,7 +24,7 @@ class BarrierOp final : public Operator<Context> {
     }
   }
 
-  virtual ~BarrierOp() {}
+  ~BarrierOp() override {}
 
   bool RunOnDevice() override {
     auto context = OperatorBase::Input<std::shared_ptr<::gloo::Context>>(0);

caffe2/contrib/gloo/broadcast_ops.h

@@ -29,7 +29,7 @@ class BroadcastOp final : public Operator<Context> {
     }
   }
 
-  virtual ~BroadcastOp() {}
+  ~BroadcastOp() override {}
 
   bool RunOnDevice() override {
     std::call_once(once_, [&] { initialize(); });

caffe2/contrib/gloo/common_world_ops.h

@@ -49,7 +49,7 @@ class CreateCommonWorld final : public Operator<Context> {
     initialize();
   }
 
-  virtual ~CreateCommonWorld() {
+  ~CreateCommonWorld() override {
   }
 
   CommonWorld rendezvousWithMPI() {
@@ -176,7 +176,7 @@ class CloneCommonWorld final : public Operator<Context> {
     }
   }
 
-  virtual ~CloneCommonWorld() {}
+  ~CloneCommonWorld() override {}
 
   bool RunOnDevice() override {
     try {

caffe2/contrib/gloo/reduce_scatter_ops.h

@@ -44,7 +44,7 @@ class ReduceScatterOp final : public Operator<Context> {
     }
   }
 
-  virtual ~ReduceScatterOp() {}
+  ~ReduceScatterOp() override {}
 
   bool RunOnDevice() override {
     std::call_once(once_, [&] { initialize(); });

caffe2/contrib/gloo/store_handler.h

@@ -12,7 +12,7 @@ class TORCH_API StoreHandlerWrapper : public ::gloo::rendezvous::Store {
  public:
   explicit StoreHandlerWrapper(StoreHandler& handler) : handler_(handler) {}
 
-  virtual ~StoreHandlerWrapper() {}
+  virtual ~StoreHandlerWrapper() override {}
 
   virtual void set(const std::string& key, const std::vector<char>& data)
       override;

caffe2/core/nomnigraph/include/nomnigraph/Representations/NeuralNet.h

@@ -168,7 +168,7 @@ class TORCH_API Tensor : public NeuralNetData {
     return D->getKind() == NNDataKind::Tensor;
   }
 
-  NeuralNetData* clone() {
+  NeuralNetData* clone() override {
     return new Tensor(name_);
   }
 
@@ -188,7 +188,7 @@ class TORCH_API Tensor : public NeuralNetData {
     name_ = name;
   }
 
-  ~Tensor() {}
+  ~Tensor() override {}
 
  private:
   std::string name_;

caffe2/core/operator.h

@@ -1157,7 +1157,7 @@ class Operator : public OperatorBase {
   template <class... Args>                                \
   explicit name(Args&&... args)                           \
       : Operator<Context>(std::forward<Args>(args)...) {} \
-  virtual ~name() noexcept {}
+  virtual ~name() noexcept override {}
 
 // Helpers to implement runtime op polymorphism. Often it's convenient to make
 // an op work on different input types (e.g. i32 vs i64 indices) or special-case

caffe2/core/qtensor_serialization.h

@@ -12,7 +12,7 @@ template <class Context>
 class QTensorSerializer : public BlobSerializerBase {
  public:
   QTensorSerializer() : context_() {}
-  ~QTensorSerializer() {}
+  ~QTensorSerializer() override {}
   /**
    * Serializes a Blob. Note that this blob has to contain QTensor<Context>.
    */

caffe2/distributed/file_store_handler.h

@@ -7,7 +7,7 @@ namespace caffe2 {
 class TORCH_API FileStoreHandler : public StoreHandler {
  public:
   explicit FileStoreHandler(const std::string& path, const std::string& prefix);
-  virtual ~FileStoreHandler();
+  ~FileStoreHandler() override;
 
   void set(const std::string& name, const std::string& data) override;
 

caffe2/ideep/operators/conv_op.cc

@@ -30,8 +30,8 @@ class IDEEPConvOp : public IDEEPConvPoolOpBase {
       algo_ = ialgo::convolution_winograd;
     }
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPConvOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPConvOp() override {}
 
   bool RunOnDeviceWithOrderNCHW() override {
     const auto& X = Input(INPUT_X);
@@ -199,8 +199,8 @@ class IDEEPConvFusionOp final : public IDEEPConvOp {
         CAFFE_THROW("Unsupported conv fusion type!");
     }
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPConvFusionOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPConvFusionOp() override {}
 };
 
 const char* kConvFusionDoc = R"DOC(
@@ -291,8 +291,8 @@ class IDEEPConvGradientOp final : public IDEEPConvPoolOpBase {
         "In order to backward propagate weights correctly, "
         "please set training_mode=1");
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPConvGradientOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPConvGradientOp() override {}
 
   bool RunOnDeviceWithOrderNCHW() override {
     const auto& X = Input(INPUT);

caffe2/ideep/operators/conv_pool_base_op.h

@@ -12,7 +12,7 @@ class IDEEPConvPoolOpBase : public ConvPoolOpBase<IDEEPContext> {
  public:
   IDEEPConvPoolOpBase(const OperatorDef& operator_def, Workspace* ws)
      : ConvPoolOpBase<IDEEPContext>(operator_def, ws) {}
-  virtual ~IDEEPConvPoolOpBase() {}
+  ~IDEEPConvPoolOpBase() override {}
 
   inline const ideep::tensor& Input(int index) {
     return OperatorBase::template Input<ideep::tensor>(index);

caffe2/ideep/operators/conv_transpose_unpool_base_op.h

@@ -116,7 +116,7 @@ class IDEEPConvTransposeUnpoolBase : public IDEEPOperator {
       CAFFE_ENFORCE_LE(adj_[dim], stride_[dim]);
     }
   }
-  virtual ~IDEEPConvTransposeUnpoolBase() {}
+  ~IDEEPConvTransposeUnpoolBase() override {}
 
   const ideep::tensor& Input(int index) {
     return OperatorBase::template Input<ideep::tensor>(index);

caffe2/ideep/operators/quantization/int8_add_op.cc

@@ -25,8 +25,8 @@ class IDEEPInt8SumReluOp final : public IDEEPOperator {
 
     Y_scales_ = ConvertScales({scale_});
   }
-  // NOLINTNEXTLINE(modernize-use-equals-default,modernize-use-override)
-  virtual ~IDEEPInt8SumReluOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPInt8SumReluOp() override {}
 
   bool RunOnDevice() override {
     itensor temp_ten;

caffe2/ideep/operators/quantization/int8_conv_op.cc

@@ -28,8 +28,8 @@ class IDEEPInt8ConvOp : public IDEEPConvPoolOpBase {
     CAFFE_ENFORCE(zero_point_ == 128 || zero_point_ == 0);
     Y_scales_ = ConvertScales({scale_});
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPInt8ConvOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPInt8ConvOp() override {}
 
   bool RunOnDeviceWithOrderNCHW() override {
     const auto &X = Input(INPUT_X);
@@ -198,8 +198,8 @@ class IDEEPInt8ConvReluOp final : public IDEEPInt8ConvOp {
     attr_ = iattr::fuse_relu();
     fusion_type_ = FUSION_CONV_RELU;
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPInt8ConvReluOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPInt8ConvReluOp() override {}
 };
 
 class IDEEPInt8ConvSumOp final : public IDEEPInt8ConvOp {
@@ -213,8 +213,8 @@ class IDEEPInt8ConvSumOp final : public IDEEPInt8ConvOp {
     attr_ = iattr::fuse_sum();
     fusion_type_ = FUSION_CONV_SUM;
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPInt8ConvSumOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPInt8ConvSumOp() override {}
 };
 
 class IDEEPInt8ConvSumReluOp final : public IDEEPInt8ConvOp {
@@ -228,8 +228,8 @@ class IDEEPInt8ConvSumReluOp final : public IDEEPInt8ConvOp {
     attr_ = iattr::residual();
     fusion_type_ = FUSION_CONV_SUM_RELU;
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPInt8ConvSumReluOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPInt8ConvSumReluOp() override {}
 };
 
 REGISTER_IDEEP_OPERATOR_WITH_ENGINE(Int8Conv, DNNLOWP, IDEEPInt8ConvOp);

caffe2/ideep/operators/quantization/int8_dequantize_op.cc

@@ -18,8 +18,8 @@ class IDEEPInt8DequantizeOp final : public IDEEPOperator {
                                               static_cast<int>(iformat::nchw)));
     }
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPInt8DequantizeOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPInt8DequantizeOp() override {}
 
   bool RunOnDevice() override {
     const auto& X = Input(0);

caffe2/ideep/operators/quantization/int8_fully_connected_op.cc

@@ -26,8 +26,8 @@ class IDEEPInt8FullyConnectedOp final : public IDEEPOperator {
     }
     Y_scales_ = ConvertScales({scale_});
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPInt8FullyConnectedOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPInt8FullyConnectedOp() override {}
 
   bool RunOnDevice() override {
     const auto& X = Input(INPUT);

caffe2/ideep/operators/quantization/int8_quantize_op.cc

@@ -26,8 +26,8 @@ class IDEEPInt8QuantizeOp final : public IDEEPOperator {
     Y_data_type_ = zero_point_ == 0 ? idtype::u8 : idtype::s8;
     Y_scales_ = ConvertScales({scale_});
   }
-  // NOLINTNEXTLINE(modernize-use-override,modernize-use-equals-default)
-  virtual ~IDEEPInt8QuantizeOp() {}
+  // NOLINTNEXTLINE(modernize-use-equals-default)
+  ~IDEEPInt8QuantizeOp() override {}
 
   bool RunOnDevice() override {
     const auto& X = Input(0);

caffe2/ideep/utils/ideep_context.h

@@ -23,20 +23,20 @@ class IDEEPContext final : public BaseContext {
 
   ~IDEEPContext() noexcept override {}
 
-  inline void SwitchToDevice(int64_t /*stream_id*/) {}
+  inline void SwitchToDevice(int64_t /*stream_id*/) override {}
   using BaseContext::SwitchToDevice;
 
-  inline void WaitEvent(const Event& ev) {
+  inline void WaitEvent(const Event& ev) override {
     ev.Wait(IDEEP, this);
   }
 
-  inline void Record(Event* ev, const char* err_msg = nullptr) const {
+  inline void Record(Event* ev, const char* err_msg = nullptr) const override {
     CAFFE_ENFORCE(ev, "Event must not be null.");
     ev->Record(IDEEP, this, err_msg);
   }
 
 
-  inline void FinishDeviceComputation() {}
+  inline void FinishDeviceComputation() override {}
 
   inline rand_gen_type& RandGenerator() {
     if (!random_generator_.get()) {

caffe2/ideep/utils/ideep_operator.h

@@ -40,7 +40,7 @@ class IDEEPOperator : public OperatorBase {
         order_(StringToStorageOrder(
             OperatorBase::GetSingleArgument<string>("order", "NCHW"))) {
   }
-  virtual ~IDEEPOperator() {}
+  ~IDEEPOperator() override {}
 
   inline const ideep::tensor& Input(int index) {
     return OperatorBase::template Input<ideep::tensor>(index);
@@ -114,7 +114,7 @@ class IDEEPOperator : public OperatorBase {
 #define USE_SIMPLE_IDEEP_CTOR_DTOR(name)                                       \
   name(const OperatorDef& operator_def, Workspace* ws)                         \
       : IDEEPOperator(operator_def, ws) {}                                     \
-  virtual ~name() {}
+  ~name() override {}
 
 // Convert zero_point scales to min_max scales
 // NOTE:

caffe2/mpi/mpi_ops.h

@@ -31,7 +31,7 @@ class MPIBroadcastOp final : public Operator<Context> {
   MPIBroadcastOp(const OperatorDef& operator_def, Workspace* ws)
       : Operator<Context>(operator_def, ws),
         root_(OperatorBase::template GetSingleArgument<int>("root", 0)) {}
-  ~MPIBroadcastOp() {}
+  ~MPIBroadcastOp() override {}
 
   bool RunOnDevice() override {
     MPI_Comm comm = OperatorBase::Input<MPICommonWorldWrapper>(0).comm();
@@ -65,7 +65,7 @@ class MPIReduceOp final : public Operator<Context> {
   MPIReduceOp(const OperatorDef& operator_def, Workspace* ws)
       : Operator<Context>(operator_def, ws),
         root_(OperatorBase::template GetSingleArgument<int>("root", 0)) {}
-  ~MPIReduceOp() {}
+  ~MPIReduceOp() override {}
 
   bool RunOnDevice() override {
     MPI_Comm comm = OperatorBase::Input<MPICommonWorldWrapper>(0).comm();

caffe2/operators/batch_gather_ops.h

@@ -50,7 +50,7 @@ class BatchGatherGradientOp final : public Operator<Context> {
       : Operator<Context>(std::forward<Args>(args)...),
         OP_SINGLE_ARG(int, "axis", axis_, 1),
         OP_SINGLE_ARG(bool, "match_outer", match_outer_, false) {}
-  virtual ~BatchGatherGradientOp() noexcept {}
+  ~BatchGatherGradientOp() noexcept override {}
 
   bool RunOnDevice() override {
     return DispatchHelper<TensorTypes<int32_t, int64_t>>::call(

caffe2/operators/batch_permutation_op.h

@@ -19,7 +19,7 @@ class BatchPermutationOp final : public Operator<Context> {
       : Operator<Context>(std::forward<Args>(args)...) {}
   USE_OPERATOR_CONTEXT_FUNCTIONS;
 
-  bool RunOnDevice();
+  bool RunOnDevice() override;
 };
 
 template <typename T, class Context>
@@ -29,7 +29,7 @@ class BatchPermutationGradientOp final : public Operator<Context> {
       : Operator<Context>(def, ws) {}
   USE_OPERATOR_CONTEXT_FUNCTIONS;
 
-  bool RunOnDevice();
+  bool RunOnDevice() override;
 };
 
 } // namespace caffe2

caffe2/operators/batch_sparse_to_dense_op.h

@@ -125,7 +125,7 @@ class BatchDenseToSparseOp : public Operator<Context> {
   template <class... Args>
   explicit BatchDenseToSparseOp(Args&&... args)
       : Operator<Context>(std::forward<Args>(args)...) {}
-  bool RunOnDevice() {
+  bool RunOnDevice() override{
     return DispatchHelper<TensorTypes<int32_t, int64_t>>::call(
         this, Input(LENGTHS));
   }

caffe2/operators/box_with_nms_limit_op.h

@@ -58,7 +58,7 @@ class BoxWithNMSLimitOp final : public Operator<Context> {
     input_scores_fg_cls_starting_id_ = (int)input_boxes_include_bg_cls_;
   }
 
-  ~BoxWithNMSLimitOp() {}
+  ~BoxWithNMSLimitOp() override {}
 
   bool RunOnDevice() override {
     if (InputSize() > 2) {

caffe2/operators/channel_backprop_stats_op.h

@@ -14,7 +14,7 @@ class ChannelBackpropStatsOp : public Operator<Context> {
   template <class... Args>
   explicit ChannelBackpropStatsOp(Args&&... args)
       : Operator<Context>(std::forward<Args>(args)...) {}
-  ~ChannelBackpropStatsOp() {}
+  ~ChannelBackpropStatsOp() override {}
 
   bool RunOnDevice() override {
     return true;