[VL] Minor fix for cpp code style (part 1) (#8157)

cpp/CMakeLists.txt

@@ -236,10 +236,7 @@ if(ENABLE_IAA)
   add_definitions(-DGLUTEN_ENABLE_IAA)
 endif()
 
-#
 # Subdirectories
-#
-
 add_subdirectory(core)
 
 if(BUILD_VELOX_BACKEND)

cpp/velox/CMakeLists.txt

@@ -140,12 +140,12 @@ endmacro()
 set(VELOX_SRCS
     compute/VeloxBackend.cc
     compute/VeloxRuntime.cc
-    compute/WholeStageResultIterator.cc
     compute/VeloxPlanConverter.cc
+    compute/WholeStageResultIterator.cc
     compute/iceberg/IcebergPlanConverter.cc
-    jni/VeloxJniWrapper.cc
     jni/JniFileSystem.cc
     jni/JniUdf.cc
+    jni/VeloxJniWrapper.cc
     memory/BufferOutputStream.cc
     memory/VeloxColumnarBatch.cc
     memory/VeloxMemoryManager.cc
@@ -154,30 +154,30 @@ set(VELOX_SRCS
     operators/functions/SparkExprToSubfieldFilterParser.cc
     operators/reader/FileReaderIterator.cc
     operators/reader/ParquetReaderIterator.cc
-    operators/serializer/VeloxColumnarToRowConverter.cc
     operators/serializer/VeloxColumnarBatchSerializer.cc
+    operators/serializer/VeloxColumnarToRowConverter.cc
     operators/serializer/VeloxRowToColumnarConverter.cc
     operators/writer/VeloxArrowWriter.cc
     operators/writer/VeloxParquetDataSource.cc
+    shuffle/VeloxHashShuffleWriter.cc
+    shuffle/VeloxRssSortShuffleWriter.cc
     shuffle/VeloxShuffleReader.cc
     shuffle/VeloxShuffleWriter.cc
-    shuffle/VeloxHashShuffleWriter.cc
     shuffle/VeloxSortShuffleWriter.cc
-    shuffle/VeloxRssSortShuffleWriter.cc
+    substrait/SubstraitExtensionCollector.cc
     substrait/SubstraitParser.cc
     substrait/SubstraitToVeloxExpr.cc
     substrait/SubstraitToVeloxPlan.cc
     substrait/SubstraitToVeloxPlanValidator.cc
     substrait/VariantToVectorConverter.cc
-    substrait/SubstraitExtensionCollector.cc
     substrait/VeloxSubstraitSignature.cc
     substrait/VeloxToSubstraitExpr.cc
     substrait/VeloxToSubstraitPlan.cc
     substrait/VeloxToSubstraitType.cc
     udf/UdfLoader.cc
-    utils/VeloxArrowUtils.cc
-    utils/ConfigExtractor.cc
     utils/Common.cc
+    utils/ConfigExtractor.cc
+    utils/VeloxArrowUtils.cc
     utils/VeloxBatchResizer.cc)
 
 if(ENABLE_S3)

cpp/velox/compute/WholeStageResultIterator.h

@@ -120,7 +120,7 @@ class WholeStageResultIterator : public ColumnarBatchIterator {
   /// All the children plan node ids with postorder traversal.
   std::vector<facebook::velox::core::PlanNodeId> orderedNodeIds_;
 
-  /// Node ids should be ommited in metrics.
+  /// Node ids should be omitted in metrics.
   std::unordered_set<facebook::velox::core::PlanNodeId> omittedNodeIds_;
   std::vector<facebook::velox::core::PlanNodeId> scanNodeIds_;
   std::vector<std::shared_ptr<SplitInfo>> scanInfos_;

cpp/velox/jni/JniFileSystem.h

@@ -24,9 +24,9 @@
 
 namespace gluten {
 
-// Register JNI-or-local (or JVM-over-local, as long as it describes what happens here)
-//   file system. maxFileSize is necessary (!= 0) because we use this size to decide
-//   whether a new file can fit in JVM heap, otherwise we write it via local fs directly.
+// Register JNI-or-local (or JVM-over-local, as long as it describes what happens here) file system. maxFileSize is
+// necessary (!= 0) because we use this size to decide whether a new file can fit in JVM heap, otherwise we write it via
+// local fs directly.
 void registerJolFileSystem(uint64_t maxFileSize);
 
 void initVeloxJniFileSystem(JNIEnv* env);

cpp/velox/memory/VeloxMemoryManager.cc

@@ -59,6 +59,7 @@ T getConfig(
   return defaultValue;
 }
 } // namespace
+
 /// We assume in a single Spark task. No thread-safety should be guaranteed.
 class ListenableArbitrator : public velox::memory::MemoryArbitrator {
  public:

cpp/velox/operators/functions/RegistrationAllFunctions.cc

@@ -43,8 +43,10 @@ void registerPrestoVectorFunctions() {
   VELOX_REGISTER_VECTOR_FUNCTION(udf_transform_values, "transform_values");
 }
 } // namespace facebook::velox::functions
+
 namespace gluten {
 namespace {
+
 void registerFunctionOverwrite() {
   velox::functions::registerUnaryNumeric<RoundFunction>({"round"});
   velox::registerFunction<RoundFunction, int8_t, int8_t, int32_t>({"round"});

cpp/velox/operators/functions/RowConstructorWithNull.cc

@@ -19,6 +19,7 @@
 #include "velox/expression/VectorFunction.h"
 
 namespace gluten {
+
 facebook::velox::TypePtr RowConstructorWithNullCallToSpecialForm::resolveType(
     const std::vector<facebook::velox::TypePtr>& argTypes) {
   auto numInput = argTypes.size();
@@ -52,4 +53,5 @@ facebook::velox::exec::ExprPtr RowConstructorWithNullCallToSpecialForm::construc
   return std::make_shared<facebook::velox::exec::Expr>(
       type, std::move(compiledChildren), function, metadata, name, trackCpuUsage);
 }
+
 } // namespace gluten

cpp/velox/operators/functions/RowConstructorWithNull.h

@@ -21,6 +21,7 @@
 #include "velox/expression/SpecialForm.h"
 
 namespace gluten {
+
 class RowConstructorWithNullCallToSpecialForm : public facebook::velox::exec::FunctionCallToSpecialForm {
  public:
   RowConstructorWithNullCallToSpecialForm(const std::string& rowFunctionName) {
@@ -49,4 +50,5 @@ class RowConstructorWithNullCallToSpecialForm : public facebook::velox::exec::Fu
  private:
   std::string rowFunctionName;
 };
+
 } // namespace gluten

cpp/velox/operators/functions/RowFunctionWithNull.h

@@ -22,10 +22,8 @@
 
 namespace gluten {
 
-/**
- * @tparam allNull If true, set struct as null when all of arguments are all, else will
- * set it null when one of its arguments is null.
- */
+///@tparam allNull If true, set struct as null when all of arguments are all, else will
+/// set it null when one of its arguments is null.
 template <bool allNull>
 class RowFunctionWithNull final : public facebook::velox::exec::VectorFunction {
  public:

cpp/velox/operators/functions/SparkExprToSubfieldFilterParser.cc

@@ -100,4 +100,5 @@ std::unique_ptr<common::Filter> SparkExprToSubfieldFilterParser::leafCallToSubfi
   }
   return nullptr;
 }
+
 } // namespace gluten

cpp/velox/operators/functions/SparkExprToSubfieldFilterParser.h

@@ -34,4 +34,5 @@ class SparkExprToSubfieldFilterParser : public facebook::velox::exec::ExprToSubf
   // Compared to the upstream 'toSubfield', the push-down of filter on nested field is disabled.
   bool toSparkSubfield(const facebook::velox::core::ITypedExpr* field, facebook::velox::common::Subfield& subfield);
 };
+
 } // namespace gluten

cpp/velox/operators/plannodes/RowVectorStream.h

@@ -24,6 +24,7 @@
 #include "velox/exec/Task.h"
 
 namespace {
+
 class SuspendedSection {
  public:
   explicit SuspendedSection(facebook::velox::exec::Driver* driver) : driver_(driver) {
@@ -42,9 +43,11 @@ class SuspendedSection {
  private:
   facebook::velox::exec::Driver* const driver_;
 };
+
 } // namespace
 
 namespace gluten {
+
 class RowVectorStream {
  public:
   explicit RowVectorStream(
@@ -67,7 +70,7 @@ class RowVectorStream {
       // possibility that this spill call hangs. See https://github.com/apache/incubator-gluten/issues/7243.
       // As of now, non-zero running threads usually happens when:
       // 1. Task A spills task B;
-      // 2. Task A trys to grow buffers created by task B, during which spill is requested on task A again.
+      // 2. Task A tries to grow buffers created by task B, during which spill is requested on task A again.
       SuspendedSection ss(driverCtx_->driver);
       hasNext = iterator_->hasNext();
     }
@@ -77,7 +80,7 @@ class RowVectorStream {
     return hasNext;
   }
 
-  // Convert arrow batch to rowvector and use new output columns
+  // Convert arrow batch to row vector and use new output columns
   facebook::velox::RowVectorPtr next() {
     if (finished_) {
       return nullptr;

cpp/velox/operators/reader/FileReaderIterator.cc

@@ -49,4 +49,5 @@ std::shared_ptr<gluten::ResultIterator> FileReaderIterator::getInputIteratorFrom
   }
   throw new GlutenException("Unreachable.");
 }
+
 } // namespace gluten

cpp/velox/operators/reader/FileReaderIterator.h

@@ -22,6 +22,7 @@
 #include "velox/common/memory/MemoryPool.h"
 
 namespace gluten {
+
 enum FileReaderType { kBuffered, kStream, kNone };
 
 class FileReaderIterator : public ColumnarBatchIterator {

cpp/velox/operators/reader/ParquetReaderIterator.cc

@@ -94,4 +94,5 @@ void ParquetBufferedReaderIterator::collectBatches() {
   auto endTime = std::chrono::steady_clock::now();
   collectBatchTime_ += std::chrono::duration_cast<std::chrono::nanoseconds>(endTime - startTime).count();
 }
-} // namespace gluten
+
+} // namespace gluten

cpp/velox/operators/serializer/VeloxColumnarBatchSerializer.cc

@@ -30,14 +30,15 @@
 using namespace facebook::velox;
 
 namespace gluten {
-
 namespace {
+
 std::unique_ptr<ByteInputStream> toByteStream(uint8_t* data, int32_t size) {
   std::vector<ByteRange> byteRanges;
   byteRanges.push_back(ByteRange{data, size, 0});
   auto byteStream = std::make_unique<BufferInputStream>(byteRanges);
   return byteStream;
 }
+
 } // namespace
 
 VeloxColumnarBatchSerializer::VeloxColumnarBatchSerializer(
@@ -89,4 +90,5 @@ std::shared_ptr<ColumnarBatch> VeloxColumnarBatchSerializer::deserialize(uint8_t
   serde_->deserialize(byteStream.get(), veloxPool_.get(), rowType_, &result, &options_);
   return std::make_shared<VeloxColumnarBatch>(result);
 }
+
 } // namespace gluten

cpp/velox/operators/serializer/VeloxRowToColumnarConverter.cc

@@ -255,6 +255,7 @@ bool supporteType(const RowTypePtr rowType) {
 }
 
 } // namespace
+
 VeloxRowToColumnarConverter::VeloxRowToColumnarConverter(
     struct ArrowSchema* cSchema,
     std::shared_ptr<memory::MemoryPool> memoryPool)
@@ -301,4 +302,5 @@ VeloxRowToColumnarConverter::convertPrimitive(int64_t numRows, int64_t* rowLengt
   auto rowVector = std::make_shared<RowVector>(pool_.get(), rowType_, BufferPtr(nullptr), numRows, std::move(columns));
   return std::make_shared<VeloxColumnarBatch>(rowVector);
 }
+
 } // namespace gluten

cpp/velox/operators/writer/VeloxArrowWriter.cc

@@ -35,4 +35,5 @@ std::shared_ptr<ColumnarBatch> VeloxArrowWriter::retrieveColumnarBatch() {
   }
   return reader_->next();
 }
+
 } // namespace gluten

cpp/velox/operators/writer/VeloxArrowWriter.h

@@ -21,6 +21,7 @@
 #include "operators/writer/ArrowWriter.h"
 
 namespace gluten {
+
 class VeloxArrowWriter : public ArrowWriter {
  public:
   explicit VeloxArrowWriter(const std::string& path, int64_t batchSize, facebook::velox::memory::MemoryPool* pool);
@@ -32,4 +33,5 @@ class VeloxArrowWriter : public ArrowWriter {
   facebook::velox::memory::MemoryPool* pool_;
   std::unique_ptr<ParquetStreamReaderIterator> reader_{nullptr};
 };
+
 } // namespace gluten

cpp/velox/operators/writer/VeloxParquetDataSource.cc

@@ -38,7 +38,6 @@ using namespace facebook::velox::common;
 using namespace facebook::velox::filesystems;
 
 namespace gluten {
-
 namespace {
 const int32_t kGzipWindowBits4k = 12;
 }

cpp/velox/operators/writer/VeloxParquetDataSourceABFS.h

@@ -51,4 +51,5 @@ class VeloxParquetDataSourceABFS final : public VeloxParquetDataSource {
         abfsFileSystem->openFileForWrite(filePath_, {{}, sinkPool_.get()}), filePath_);
   }
 };
+
 } // namespace gluten

cpp/velox/operators/writer/VeloxParquetDataSourceGCS.h

@@ -32,6 +32,7 @@
 #include "velox/dwio/common/Options.h"
 
 namespace gluten {
+
 class VeloxParquetDataSourceGCS final : public VeloxParquetDataSource {
  public:
   VeloxParquetDataSourceGCS(
@@ -48,4 +49,5 @@ class VeloxParquetDataSourceGCS final : public VeloxParquetDataSource {
         gcsFileSystem->openFileForWrite(filePath_, {{}, sinkPool_.get()}), filePath_);
   }
 };
+
 } // namespace gluten

cpp/velox/operators/writer/VeloxParquetDataSourceHDFS.h

@@ -48,4 +48,5 @@ class VeloxParquetDataSourceHDFS final : public VeloxParquetDataSource {
     sink_ = dwio::common::FileSink::create(filePath_, {.connectorProperties = hiveConf, .pool = sinkPool_.get()});
   }
 };
+
 } // namespace gluten

cpp/velox/operators/writer/VeloxParquetDataSourceS3.h

@@ -48,4 +48,5 @@ class VeloxParquetDataSourceS3 final : public VeloxParquetDataSource {
     sink_ = dwio::common::FileSink::create(filePath_, {.connectorProperties = hiveConf, .pool = sinkPool_.get()});
   }
 };
+
 } // namespace gluten

cpp/velox/shuffle/GlutenByteStream.h

@@ -19,6 +19,7 @@
 #include "velox/common/memory/ByteStream.h"
 
 namespace facebook::velox {
+
 class GlutenByteInputStream : public ByteInputStream {
  protected:
   /// TODO Remove after refactoring SpillInput.
@@ -264,4 +265,5 @@ inline int128_t GlutenByteInputStream::read<int128_t>() {
   readBytes(reinterpret_cast<uint8_t*>(&value), sizeof(value));
   return value;
 }
+
 } // namespace facebook::velox