refactor : Change C-style casts to C++-style (Part 3)

As per the security guideline in
https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#es49-if-you-must-use-a-cast-use-a-named-cast

Covers the findings in velox/vector

velox/vector/VectorSaver.cpp

@@ -36,7 +36,7 @@ enum class Encoding : int8_t {
 
 template <typename T>
 void write(const T& value, std::ostream& out) {
-  out.write((char*)&value, sizeof(T));
+  out.write(reinterpret_cast<const char*>(&value), sizeof(T));
 }
 
 template <>
@@ -53,7 +53,7 @@ void write<std::string>(const std::string& value, std::ostream& out) {
 template <typename T>
 T read(std::istream& in) {
   T value;
-  in.read((char*)&value, sizeof(T));
+  in.read(reinterpret_cast<char*>(&value), sizeof(T));
   return value;
 }
 
@@ -78,16 +78,16 @@ void writeEncoding(VectorEncoding::Simple encoding, std::ostream& out) {
     case VectorEncoding::Simple::ROW:
     case VectorEncoding::Simple::ARRAY:
     case VectorEncoding::Simple::MAP:
-      write<int32_t>((int8_t)Encoding::kFlat, out);
+      write<int32_t>(static_cast<int8_t>(Encoding::kFlat), out);
       return;
     case VectorEncoding::Simple::CONSTANT:
-      write<int32_t>((int8_t)Encoding::kConstant, out);
+      write<int32_t>(static_cast<int8_t>(Encoding::kConstant), out);
       return;
     case VectorEncoding::Simple::DICTIONARY:
-      write<int32_t>((int8_t)Encoding::kDictionary, out);
+      write<int32_t>(static_cast<int8_t>(Encoding::kDictionary), out);
       return;
     case VectorEncoding::Simple::LAZY:
-      write<int32_t>((int8_t)Encoding::kLazy, out);
+      write<int32_t>(static_cast<int8_t>(Encoding::kLazy), out);
       return;
     default:
       VELOX_UNSUPPORTED("Unsupported encoding: {}", mapSimpleToName(encoding));
@@ -275,7 +275,7 @@ void writeScalarConstant(const BaseVector& vector, std::ostream& out) {
   using T = typename TypeTraits<kind>::NativeType;
 
   auto value = vector.as<ConstantVector<T>>()->valueAt(0);
-  out.write((const char*)&value, sizeof(T));
+  out.write(reinterpret_cast<const char*>(&value), sizeof(T));
 
   if constexpr (std::is_same_v<T, StringView>) {
     if (!value.isInline()) {

velox/vector/VectorStream.cpp

@@ -323,7 +323,9 @@ RowVectorPtr IOBufToRowVector(
 
   for (const auto& range : ioBuf) {
     ranges.emplace_back(ByteRange{
-        const_cast<uint8_t*>(range.data()), (int32_t)range.size(), 0});
+        const_cast<uint8_t*>(range.data()),
+        static_cast<int32_t>(range.size()),
+        0});
   }
 
   auto byteStream = std::make_unique<BufferInputStream>(std::move(ranges));

velox/vector/arrow/Bridge.cpp

@@ -72,7 +72,7 @@ class VeloxToArrowBridgeHolder {
   }
 
   const void** getArrowBuffers() {
-    return (const void**)&(buffers_[0]);
+    return reinterpret_cast<const void**>(&(buffers_[0]));
   }
 
   // Allocates space for `numChildren` ArrowArray pointers.
@@ -564,7 +564,8 @@ VectorPtr createStringFlatVectorFromUtf8View(
       "Expecting three or more buffers as input for string view types.");
 
   // The last C data buffer stores buffer sizes
-  auto* bufferSizes = (uint64_t*)arrowArray.buffers[num_buffers - 1];
+  auto* bufferSizes =
+      reinterpret_cast<const uint64_t*>(arrowArray.buffers[num_buffers - 1]);
   std::vector<BufferPtr> stringViewBuffers(num_buffers - 3);
 
   // Skipping buffer_id = 0 (nulls buffer) and buffer_id = 1 (values buffer)
@@ -582,13 +583,15 @@ VectorPtr createStringFlatVectorFromUtf8View(
   // buffer-index, 4-byte buffer-offset] to 16-byte Velox StringView [4-byte
   // length, 4-byte prefix, 8-byte buffer-ptr]
   for (int32_t idx_64 = 0; idx_64 < arrowArray.length; ++idx_64) {
-    auto* view = (uint32_t*)(&((uint64_t*)arrowArray.buffers[1])[2 * idx_64]);
-    rawStringViews[2 * idx_64] = *(uint64_t*)view;
+    auto* view = reinterpret_cast<const uint32_t*>(&(
+        reinterpret_cast<const uint64_t*>(arrowArray.buffers[1]))[2 * idx_64]);
+    rawStringViews[2 * idx_64] = *reinterpret_cast<const uint64_t*>(view);
     if (view[0] > 12)
       rawStringViews[2 * idx_64 + 1] =
-          (uint64_t)arrowArray.buffers[2 + view[2]] + view[3];
+          reinterpret_cast<uint64_t>(arrowArray.buffers[2 + view[2]]) + view[3];
     else
-      rawStringViews[2 * idx_64 + 1] = *(uint64_t*)&view[2];
+      rawStringViews[2 * idx_64 + 1] =
+          *reinterpret_cast<const uint64_t*>(&view[2]);
   }
 
   return std::make_shared<FlatVector<StringView>>(
@@ -777,18 +780,18 @@ void exportViews(
       stringBufferVec.begin(),
       stringBufferVec.end(),
       [&out](const auto& lhs, const auto& rhs) {
-        return ((uint64_t*)&out.buffers[2])[lhs] <
-            ((uint64_t*)&out.buffers[2])[rhs];
+        return reinterpret_cast<uint64_t*>(&out.buffers[2])[lhs] <
+            reinterpret_cast<uint64_t*>(&out.buffers[2])[rhs];
       });
 
-  auto utf8Views = (uint64_t*)out.buffers[1];
+  auto utf8Views = (uint64_t*)(out.buffers[1]);
   int32_t bufferIdxCache = 0;
   uint64_t bufferAddrCache = 0;
 
   rows.apply([&](vector_size_t i) {
-    auto view = (uint32_t*)&utf8Views[2 * i];
+    auto view = reinterpret_cast<uint32_t*>(&utf8Views[2 * i]);
     if (!vec.isNullAt(i) && view[0] > 12) {
-      uint64_t currAddr = *(uint64_t*)&view[2];
+      uint64_t currAddr = *reinterpret_cast<const uint64_t*>(&view[2]);
       // 2. Search for correct index with the buffer-pointer as key. Cache the
       // found buffer's address and index in bufferAddrCache and bufferIdxCache
       // respectively
@@ -800,9 +803,9 @@ void exportViews(
             stringBufferVec.end(),
             currAddr,
             [&out](const auto& lhs, const auto& rhs) {
-              return lhs < ((uint64_t*)&out.buffers[2])[rhs];
+              return lhs < (reinterpret_cast<uint64_t*>(&out.buffers[2]))[rhs];
             }));
-        bufferAddrCache = ((uint64_t*)&out.buffers[2])[*it];
+        bufferAddrCache = (reinterpret_cast<uint64_t*>(&out.buffers[2]))[*it];
         bufferIdxCache = *it;
       }
       view[2] = bufferIdxCache;

velox/vector/fuzzer/examples/EncodingGeneratorExample.cpp

@@ -67,7 +67,9 @@ int main() {
   }
 
   std::cout << "Probability of constant encoding = "
-            << (double)numConst / ((double)numConst + (double)numDict) << "\n";
+            << static_cast<double>(numConst) /
+          static_cast<double>(numConst + numDict)
+            << "\n";
 
   return 0;
 }