chore: update pre-commit cmake format version and add cpplint check

lucasfang · lucasfang · commit ec88547d7a7f · 2025-12-17T01:30:33.000Z
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -43,7 +43,7 @@ repos:
         exclude: (^test/test_data/.*|^third_party/.*)
 
   - repo: https://github.com/cheshirekow/cmake-format-precommit
-    rev: v0.6.10
+    rev: v0.6.13
     hooks:
       - id: cmake-format
         exclude: (^test/test_data/.*|^third_party/.*)
@@ -67,3 +67,23 @@ repos:
           '--disable',
           'dangling-hyphen,line-too-long',
         ]
+
+  - repo: https://github.com/cpplint/cpplint
+    rev: 1.6.1
+    hooks:
+      - id: cpplint
+        alias: cpp
+        name: C++ Lint
+        args:
+          - "--quiet"
+          - "--verbose=2"
+          - "--filter=-whitespace/line_length,-whitespace/parens,-build/c++11,-readability/nolint,-runtime/references"
+
+        types_or:
+          - c++
+        exclude: |
+          (?x)^(
+            (third_party)/.*|
+            (test)/.*|
+            (src)/.*
+          )$
diff --git a/cmake_modules/BuildUtils.cmake b/cmake_modules/BuildUtils.cmake
@@ -291,7 +291,9 @@ function(add_test_case REL_TEST_NAME)
         add_dependencies(${LABEL_TEST_NAME} ${TEST_NAME})
     endforeach()
 
-    set_property(TEST ${TEST_NAME} APPEND PROPERTY LABELS ${LABELS})
+    set_property(TEST ${TEST_NAME}
+                 APPEND
+                 PROPERTY LABELS ${LABELS})
 endfunction()
 
 # Adding unit tests part of the "paimon" portion of the test suite
diff --git a/cmake_modules/SetupCxxFlags.cmake b/cmake_modules/SetupCxxFlags.cmake
@@ -297,8 +297,8 @@ if(NOT WIN32 AND NOT APPLE)
         # won't find it when invoked directly from make/ninja (which is typically
         # done outside devtoolset).
         execute_process(COMMAND which ld.gold
-                        OUTPUT_VARIABLE GOLD_LOCATION OUTPUT_STRIP_TRAILING_WHITESPACE
-                                        ERROR_QUIET)
+                        OUTPUT_VARIABLE GOLD_LOCATION
+                        OUTPUT_STRIP_TRAILING_WHITESPACE ERROR_QUIET)
         if("${GOLD_LOCATION}" MATCHES "^/opt/rh/devtoolset")
             message("Skipping optional gold linker (version ${GOLD_VERSION}) because "
                     "it's in devtoolset")
diff --git a/cmake_modules/ThirdpartyToolchain.cmake b/cmake_modules/ThirdpartyToolchain.cmake
@@ -251,8 +251,7 @@ macro(build_fmt)
     externalproject_add(fmt_ep
                         URL ${FMT_SOURCE_URL}
                         URL_HASH "SHA256=${PAIMON_FMT_BUILD_SHA256_CHECKSUM}"
-                        ${FMT_CONFIGURE}
-                        ${THIRDPARTY_LOG_OPTIONS}
+                        ${FMT_CONFIGURE} ${THIRDPARTY_LOG_OPTIONS}
                         BUILD_BYPRODUCTS ${FMT_STATIC_LIB})
 
     file(MAKE_DIRECTORY "${FMT_INCLUDE_DIR}")
@@ -340,8 +339,7 @@ macro(build_zstd)
     externalproject_add(zstd_ep
                         URL ${ZSTD_SOURCE_URL}
                         URL_HASH "SHA256=${PAIMON_ZSTD_BUILD_SHA256_CHECKSUM}"
-                        ${ZSTD_CONFIGURE}
-                        ${THIRDPARTY_LOG_OPTIONS}
+                        ${ZSTD_CONFIGURE} ${THIRDPARTY_LOG_OPTIONS}
                         BUILD_BYPRODUCTS ${ZSTD_STATIC_LIB})
 
     file(MAKE_DIRECTORY "${ZSTD_INCLUDE_DIR}")
@@ -367,8 +365,7 @@ macro(build_lz4)
     externalproject_add(lz4_ep
                         URL ${LZ4_SOURCE_URL}
                         URL_HASH "SHA256=${PAIMON_LZ4_BUILD_SHA256_CHECKSUM}"
-                        ${LZ4_CONFIGURE}
-                        ${THIRDPARTY_LOG_OPTIONS}
+                        ${LZ4_CONFIGURE} ${THIRDPARTY_LOG_OPTIONS}
                         BUILD_BYPRODUCTS ${LZ4_STATIC_LIB})
 
     file(MAKE_DIRECTORY "${LZ4_INCLUDE_DIR}")
@@ -510,8 +507,7 @@ macro(build_protobuf)
     externalproject_add(protobuf_ep
                         URL ${PROTOBUF_SOURCE_URL}
                         URL_HASH "SHA256=${PAIMON_PROTOBUF_BUILD_SHA256_CHECKSUM}"
-                        ${PROTOBUF_CONFIGURE}
-                        ${THIRDPARTY_LOG_OPTIONS}
+                        ${PROTOBUF_CONFIGURE} ${THIRDPARTY_LOG_OPTIONS}
                         # BUILD_IN_SOURCE 1
                         BUILD_BYPRODUCTS "${PROTOBUF_STATIC_LIB}" "${PROTOBUF_COMPILER}"
                         DEPENDS zlib)
diff --git a/include/paimon/file_index/file_indexer.h b/include/paimon/file_index/file_indexer.h
@@ -16,6 +16,8 @@
 
 #pragma once
 
+#include <memory>
+
 #include "paimon/file_index/file_index_reader.h"
 #include "paimon/file_index/file_index_result.h"
 #include "paimon/file_index/file_index_writer.h"
diff --git a/include/paimon/format/file_format.h b/include/paimon/format/file_format.h
@@ -16,6 +16,9 @@
 
 #pragma once
 
+#include <memory>
+#include <string>
+
 #include "paimon/format/format_stats_extractor.h"
 #include "paimon/format/reader_builder.h"
 #include "paimon/format/writer_builder.h"
diff --git a/include/paimon/format/format_stats_extractor.h b/include/paimon/format/format_stats_extractor.h
@@ -18,6 +18,7 @@
 
 #include <memory>
 #include <string>
+#include <utility>
 #include <vector>
 
 #include "paimon/result.h"
diff --git a/include/paimon/format/reader_builder.h b/include/paimon/format/reader_builder.h
@@ -17,6 +17,7 @@
 #pragma once
 
 #include <memory>
+#include <string>
 
 #include "paimon/memory/memory_pool.h"
 #include "paimon/reader/file_batch_reader.h"
diff --git a/include/paimon/global_index/bitmap_global_index_result.h b/include/paimon/global_index/bitmap_global_index_result.h
@@ -18,6 +18,7 @@
 
 #include <memory>
 #include <string>
+#include <utility>
 
 #include "paimon/global_index/global_index_result.h"
 #include "paimon/utils/range.h"
@@ -33,7 +34,8 @@ namespace paimon {
 class PAIMON_EXPORT BitmapGlobalIndexResult : public GlobalIndexResult {
  public:
     using BitmapSupplier = std::function<Result<RoaringBitmap64>()>;
-    BitmapGlobalIndexResult(BitmapSupplier bitmap_supplier) : bitmap_supplier_(bitmap_supplier) {}
+    explicit BitmapGlobalIndexResult(BitmapSupplier bitmap_supplier)
+        : bitmap_supplier_(bitmap_supplier) {}
 
     class Iterator : public GlobalIndexResult::Iterator {
      public:
diff --git a/include/paimon/global_index/bitmap_topk_global_index_result.h b/include/paimon/global_index/bitmap_topk_global_index_result.h
@@ -18,6 +18,8 @@
 
 #include <memory>
 #include <string>
+#include <utility>
+#include <vector>
 
 #include "paimon/global_index/global_index_result.h"
 #include "paimon/utils/roaring_bitmap64.h"
@@ -35,7 +37,7 @@ class PAIMON_EXPORT BitmapTopKGlobalIndexResult : public TopKGlobalIndexResult {
  public:
     BitmapTopKGlobalIndexResult(RoaringBitmap64&& bitmap, std::vector<float>&& scores)
         : bitmap_(std::move(bitmap)), scores_(std::move(scores)) {
-        assert((size_t)bitmap_.Cardinality() == scores_.size());
+        assert(static_cast<size_t>(bitmap_.Cardinality()) == scores_.size());
     }
 
     class TopKIterator : public TopKGlobalIndexResult::TopKIterator {
diff --git a/include/paimon/global_index/global_index_evaluator.h b/include/paimon/global_index/global_index_evaluator.h
@@ -15,6 +15,9 @@
  */
 
 #pragma once
+
+#include <memory>
+
 #include "paimon/global_index/global_index_result.h"
 #include "paimon/predicate/predicate.h"
 #include "paimon/visibility.h"
diff --git a/include/paimon/global_index/global_index_io_meta.h b/include/paimon/global_index/global_index_io_meta.h
@@ -16,6 +16,9 @@
 
 #pragma once
 
+#include <memory>
+#include <string>
+
 #include "paimon/memory/bytes.h"
 #include "paimon/utils/range.h"
 
diff --git a/include/paimon/global_index/global_index_reader.h b/include/paimon/global_index/global_index_reader.h
@@ -15,7 +15,10 @@
  */
 
 #pragma once
+
 #include <functional>
+#include <memory>
+#include <vector>
 
 #include "paimon/global_index/global_index_result.h"
 #include "paimon/predicate/function_visitor.h"
diff --git a/include/paimon/global_index/global_index_result.h b/include/paimon/global_index/global_index_result.h
@@ -18,6 +18,7 @@
 
 #include <memory>
 #include <string>
+#include <utility>
 
 #include "paimon/result.h"
 #include "paimon/visibility.h"
diff --git a/include/paimon/global_index/global_index_scan.h b/include/paimon/global_index/global_index_scan.h
@@ -15,8 +15,13 @@
  */
 
 #pragma once
+
+#include <map>
+#include <memory>
 #include <optional>
 #include <set>
+#include <string>
+#include <vector>
 
 #include "paimon/global_index/row_range_global_index_scanner.h"
 #include "paimon/utils/range.h"
diff --git a/include/paimon/global_index/global_index_writer.h b/include/paimon/global_index/global_index_writer.h
@@ -16,6 +16,8 @@
 
 #pragma once
 
+#include <vector>
+
 #include "paimon/global_index/global_index_io_meta.h"
 #include "paimon/result.h"
 #include "paimon/status.h"
diff --git a/include/paimon/global_index/global_indexer.h b/include/paimon/global_index/global_indexer.h
@@ -16,6 +16,10 @@
 
 #pragma once
 
+#include <memory>
+#include <string>
+#include <vector>
+
 #include "paimon/global_index/global_index_io_meta.h"
 #include "paimon/global_index/global_index_reader.h"
 #include "paimon/global_index/global_index_writer.h"
diff --git a/include/paimon/global_index/io/global_index_file_reader.h b/include/paimon/global_index/io/global_index_file_reader.h
@@ -16,6 +16,7 @@
 
 #pragma once
 
+#include <memory>
 #include <string>
 
 #include "paimon/result.h"
diff --git a/include/paimon/global_index/io/global_index_file_writer.h b/include/paimon/global_index/io/global_index_file_writer.h
@@ -16,6 +16,7 @@
 
 #pragma once
 
+#include <memory>
 #include <string>
 
 #include "paimon/fs/file_system.h"
diff --git a/include/paimon/global_index/row_range_global_index_scanner.h b/include/paimon/global_index/row_range_global_index_scanner.h
@@ -15,6 +15,10 @@
  */
 
 #pragma once
+
+#include <memory>
+#include <string>
+
 #include "paimon/global_index/global_index_evaluator.h"
 #include "paimon/global_index/global_index_reader.h"
 #include "paimon/visibility.h"
diff --git a/include/paimon/global_index/row_range_global_index_writer.h b/include/paimon/global_index/row_range_global_index_writer.h
@@ -15,7 +15,9 @@
  */
 
 #pragma once
+
 #include <map>
+#include <memory>
 #include <string>
 
 #include "paimon/memory/memory_pool.h"
diff --git a/include/paimon/macros.h b/include/paimon/macros.h
@@ -120,7 +120,7 @@
 
 // macros to disable padding
 // these macros are portable across different compilers and platforms
-//[https://github.com/google/flatbuffers/blob/master/include/flatbuffers/flatbuffers.h#L1355]
+// [https://github.com/google/flatbuffers/blob/master/include/flatbuffers/flatbuffers.h#L1355]
 #if !defined(MANUALLY_ALIGNED_STRUCT)
 #if defined(_MSC_VER)
 #define MANUALLY_ALIGNED_STRUCT(alignment) \
diff --git a/include/paimon/predicate/function_visitor.h b/include/paimon/predicate/function_visitor.h
@@ -17,6 +17,7 @@
 #pragma once
 
 #include <string>
+#include <vector>
 
 #include "paimon/predicate/leaf_predicate.h"
 #include "paimon/predicate/literal.h"
diff --git a/include/paimon/reader/file_batch_reader.h b/include/paimon/reader/file_batch_reader.h
@@ -16,6 +16,8 @@
 
 #pragma once
 
+#include <memory>
+#include <utility>
 #include <vector>
 
 #include "paimon/predicate/predicate.h"
diff --git a/include/paimon/result.h b/include/paimon/result.h
@@ -18,6 +18,7 @@
 
 #include <cassert>
 #include <type_traits>
+#include <utility>
 
 #include "paimon/status.h"
 #include "paimon/traits.h"
@@ -38,11 +39,11 @@ class PAIMON_MUST_USE_TYPE PAIMON_EXPORT Result {
 
     /// Construct a successful result with a copy of the given value.
     /// @param data The value to store in result.
-    Result(const T& data) : status_(), data_(data) {}  // NOLINT(google-explicit-constructor)
+    Result(const T& data) : status_(), data_(data) {}  // NOLINT(runtime/explicit)
 
     /// Construct a successful result by moving the given value.
     /// @param data The value to move into result.
-    Result(T&& data) : status_(), data_(std::move(data)) {}  // NOLINT(google-explicit-constructor)
+    Result(T&& data) : status_(), data_(std::move(data)) {}  // NOLINT(runtime/explicit)
 
     /// Template constructor for converting compatible pointer types.
     /// Support T = std::unique_ptr<B> and U = std::unique_ptr<D> convert, where D is derived class
@@ -52,11 +53,11 @@ class PAIMON_MUST_USE_TYPE PAIMON_EXPORT Result {
               std::enable_if_t<
                   is_pointer<U>::value && is_pointer<T>::value &&
                   std::is_convertible_v<value_type_traits_t<U>, value_type_traits_t<T>>>* = nullptr>
-    Result(U&& data) : status_(), data_(std::move(data)) {}  // NOLINT(google-explicit-constructor)
+    Result(U&& data) : status_(), data_(std::move(data)) {}  // NOLINT(runtime/explicit)
 
     /// Construct a failed result with the given status.
     /// @param status The status object describing the error.
-    Result(const Status& status) : status_(status) {}  // NOLINT(google-explicit-constructor)
+    Result(const Status& status) : status_(status) {}  // NOLINT(runtime/explicit)
 
     /// Copy constructor.
     /// @param other The result to copy from.
@@ -84,7 +85,7 @@ class PAIMON_MUST_USE_TYPE PAIMON_EXPORT Result {
               std::enable_if_t<
                   is_pointer<U>::value && is_pointer<T>::value &&
                   std::is_convertible_v<value_type_traits_t<U>, value_type_traits_t<T>>>* = nullptr>
-    Result(Result<U>&& other) noexcept {  // NOLINT(google-explicit-constructor)
+    Result(Result<U>&& other) noexcept {  // NOLINT(runtime/explicit)
         if (other.ok()) {
             MakeValue(std::move(other).value());
         }
diff --git a/src/paimon/common/data/binary_data_read_utils.h b/src/paimon/common/data/binary_data_read_utils.h
@@ -16,6 +16,9 @@
 
 #pragma once
 
+#include <memory>
+#include <vector>
+
 #include "paimon/common/data/binary_array.h"
 #include "paimon/common/data/binary_row.h"
 #include "paimon/common/data/binary_string.h"
diff --git a/src/paimon/common/predicate/predicate_validator_test.cpp b/src/paimon/common/predicate/predicate_validator_test.cpp
diff --git a/src/paimon/core/mergetree/merge_tree_writer_test.cpp b/src/paimon/core/mergetree/merge_tree_writer_test.cpp
diff --git a/src/paimon/format/lance/lance_reader_builder.h b/src/paimon/format/lance/lance_reader_builder.h
diff --git a/src/paimon/format/orc/orc_file_batch_reader.cpp b/src/paimon/format/orc/orc_file_batch_reader.cpp
diff --git a/src/paimon/global_index/lumina/lumina_file_writer.h b/src/paimon/global_index/lumina/lumina_file_writer.h
