[ci] run `include-what-you-use` in CI

[jameslamb]

Summary

In early 2024, the R package was almost archived on CRAN because of a compiler warning from clang-18 resulting from an unnecessary #include <string>.

ref: #6264

This project should test running include-what-you-use (https://github.com/include-what-you-use/include-what-you-use) in CI to catch such things earlier.

Motivation

Removing unused includes in C/C++ code could have the following benefits:

• smaller binary size
• reduced risk of runtime failures caused by irrelevant ABI breakages
• reducing build / linking times
• reduced effort required to update to new C++ standards

Description

Consider adding that to the cpp_tests CI jobs, and maybe to the R-package CMake CI jobs.

Try reverting the change from #6265 and test if include-what-you-use would have caught it.

References

See the documentation at https://github.com/include-what-you-use/include-what-you-use?tab=readme-ov-file#using-with-cmake.

[jameslamb]

I tried this on my mac tonight.

Installed with Homebrew.

brew install include-what-you-use

Then ran the following (compiler = Apple clang version 13.0.0 (clang-1300.0.27.3)).

cmake \
    -B build \
    -S . \
    -DCMAKE_C_INCLUDE_WHAT_YOU_USE=include-what-you-use \
    -DCMAKE_CXX_INCLUDE_WHAT_YOU_USE=include-what-you-use

cmake \
    --build build \
    --target _lightgbm \
    2>&1 \
| tee ./logs.txt \
| cat

That generated a lot of recommendations.

logs.txt (click me)

[  2%] Building CXX object CMakeFiles/lightgbm_capi_objs.dir/src/c_api.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/c_api.h should add these lines:
struct ArrowArray;
struct ArrowSchema;

/Users/jlamb/repos/LightGBM/include/LightGBM/c_api.h should remove these lines:
- #include <LightGBM/arrow.h>  // lines 16-16
- #include <cstring>  // lines 22-22

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/c_api.h:
#include <LightGBM/export.h>  // for LIGHTGBM_C_EXPORT
#include <cstdint>            // for int64_t, int32_t, uint8_t
#include <cstdio>             // for size_t, snprintf
struct ArrowArray;
struct ArrowSchema;
---

/Users/jlamb/repos/LightGBM/src/c_api.cpp should add these lines:
#include <__exception/exception.h>                         // for exception
#include <__mutex/unique_lock.h>                           // for unique_lock
#include <algorithm>                                       // for max, min
#include <cmath>                                           // for fabs, isnan
#include <cstring>                                         // for memcpy
#include <iterator>                                        // for advance
#include <unordered_map>                                   // for unordered_map
#include <utility>                                         // for pair, make...
#include "LightGBM/arrow.tpp"                              // for ArrowChunk...
#include "LightGBM/bin.h"                                  // for BinMapper
#include "LightGBM/meta.h"                                 // for kZeroThres...
#include "omp.h"                                           // for omp_get_th...

/Users/jlamb/repos/LightGBM/src/c_api.cpp should remove these lines:
- #include <LightGBM/prediction_early_stop.h>  // lines 15-15
- #include <LightGBM/utils/threading.h>  // lines 21-21
- #include <mutex>  // lines 27-27
- #include <stdexcept>  // lines 28-28

The full include-list for /Users/jlamb/repos/LightGBM/src/c_api.cpp:
#include <LightGBM/c_api.h>
#include <LightGBM/arrow.h>                                // for ArrowTable
#include <LightGBM/boosting.h>                             // for Boosting
#include <LightGBM/config.h>                               // for Config
#include <LightGBM/dataset.h>                              // for Dataset
#include <LightGBM/dataset_loader.h>                       // for DatasetLoader
#include <LightGBM/metric.h>                               // for Metric
#include <LightGBM/network.h>                              // for Network
#include <LightGBM/objective_function.h>                   // for ObjectiveF...
#include <LightGBM/utils/byte_buffer.h>                    // for ByteBuffer
#include <LightGBM/utils/common.h>                         // for Vector2Ptr
#include <LightGBM/utils/log.h>                            // for Log, CHECK_LT
#include <LightGBM/utils/openmp_wrapper.h>                 // for OMP_SET_NU...
#include <LightGBM/utils/random.h>                         // for Random
#include <__exception/exception.h>                         // for exception
#include <__mutex/unique_lock.h>                           // for unique_lock
#include <LightGBM/utils/yamc/alternate_shared_mutex.hpp>  // for shared_mutex
#include <LightGBM/utils/yamc/yamc_shared_lock.hpp>        // for shared_lock
#include <algorithm>                                       // for max, min
#include <cmath>                                           // for fabs, isnan
#include <cstdio>                                          // for size_t
#include <cstring>                                         // for memcpy
#include <functional>                                      // for function
#include <iterator>                                        // for advance
#include <memory>                                          // for unique_ptr
#include <string>                                          // for basic_string
#include <unordered_map>                                   // for unordered_map
#include <utility>                                         // for pair, make...
#include <vector>                                          // for vector
#include "LightGBM/arrow.tpp"                              // for ArrowChunk...
#include "LightGBM/bin.h"                                  // for BinMapper
#include "LightGBM/meta.h"                                 // for kZeroThres...
#include "application/predictor.hpp"                       // for Predictor
#include "omp.h"                                           // for omp_get_th...
---

[  2%] Built target lightgbm_capi_objs
[  5%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/boosting.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/boosting.h should add these lines:
#include <stddef.h>           // for size_t
#include <stdint.h>           // for int64_t
#include "LightGBM/export.h"  // for LIGHTGBM_EXPORT
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/include/LightGBM/boosting.h should remove these lines:
- #include <LightGBM/config.h>  // lines 8-8
- #include <map>  // lines 12-12

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/boosting.h:
#include <LightGBM/meta.h>    // for score_t
#include <stddef.h>           // for size_t
#include <stdint.h>           // for int64_t
#include <string>             // for string
#include <unordered_map>      // for unordered_map
#include <vector>             // for vector
#include "LightGBM/export.h"  // for LIGHTGBM_EXPORT
namespace LightGBM { class Dataset; }  // lines 19-19
namespace LightGBM { class Metric; }  // lines 21-21
namespace LightGBM { class ObjectiveFunction; }  // lines 20-20
namespace LightGBM { struct Config; }
namespace LightGBM { struct PredictionEarlyStopInstance; }  // lines 22-22
---

/Users/jlamb/repos/LightGBM/src/boosting/boosting.cpp should add these lines:
#include <chrono>                        // for duration, operator*, operator-
#include <memory>                        // for unique_ptr
#include <ratio>                         // for ratio, milli
#include "LightGBM/utils/log.h"          // for Log
#include "LightGBM/utils/text_reader.h"  // for TextReader

/Users/jlamb/repos/LightGBM/src/boosting/boosting.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/boosting/boosting.cpp:
#include <LightGBM/boosting.h>
#include <chrono>                        // for duration, operator*, operator-
#include <memory>                        // for unique_ptr
#include <ratio>                         // for ratio, milli
#include "LightGBM/utils/log.h"          // for Log
#include "LightGBM/utils/text_reader.h"  // for TextReader
#include "dart.hpp"                      // for DART
#include "gbdt.h"                        // for GBDT
#include "rf.hpp"                        // for RF
---

[  8%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/boosting/gbdt.h should add these lines:
#include <__fwd/sstream.h>                  // for stringstream
#include <stdint.h>                         // for int64_t, int8_t
#include <sstream>                          // for operator<<, basic_ostream
#include <type_traits>                      // for __libcpp_remove_reference_t
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for score_t, data_size_t, kAl...
#include "LightGBM/tree.h"                  // for Tree
#include "LightGBM/tree_learner.h"          // for TreeLearner
#include "LightGBM/utils/common.h"          // for Split, AlignmentAllocator
#include "LightGBM/utils/log.h"             // for CHECK, Log
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
#include "LightGBM/utils/random.h"          // for Random
namespace LightGBM { class Metric; }
namespace LightGBM { struct PredictionEarlyStopInstance; }

/Users/jlamb/repos/LightGBM/src/boosting/gbdt.h should remove these lines:
- #include <LightGBM/cuda/vector_cudahost.h>  // lines 11-11
- #include <LightGBM/prediction_early_stop.h>  // lines 10-10
- #include <LightGBM/utils/threading.h>  // lines 13-13
- #include <fstream>  // lines 19-19
- #include <map>  // lines 20-20
- #include <mutex>  // lines 22-22
- #include "cuda/cuda_score_updater.hpp"  // lines 27-27

The full include-list for /Users/jlamb/repos/LightGBM/src/boosting/gbdt.h:
#include <LightGBM/boosting.h>              // for GBDTBase
#include <LightGBM/objective_function.h>    // for ObjectiveFunction
#include <LightGBM/sample_strategy.h>       // for SampleStrategy
#include <LightGBM/utils/json11.h>          // for Json
#include <__fwd/sstream.h>                  // for stringstream
#include <stdint.h>                         // for int64_t, int8_t
#include <algorithm>                        // for min, max
#include <cstdio>                           // for size_t
#include <memory>                           // for unique_ptr, allocator
#include <sstream>                          // for operator<<, basic_ostream
#include <string>                           // for basic_string, char_traits
#include <type_traits>                      // for __libcpp_remove_reference_t
#include <unordered_map>                    // for unordered_map, __hash_map...
#include <utility>                          // for move, swap, pair
#include <vector>                           // for vector
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for score_t, data_size_t, kAl...
#include "LightGBM/tree.h"                  // for Tree
#include "LightGBM/tree_learner.h"          // for TreeLearner
#include "LightGBM/utils/common.h"          // for Split, AlignmentAllocator
#include "LightGBM/utils/log.h"             // for CHECK, Log
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
#include "LightGBM/utils/random.h"          // for Random
#include "score_updater.hpp"                // for ScoreUpdater
namespace LightGBM { class Metric; }
namespace LightGBM { struct PredictionEarlyStopInstance; }
---

/Users/jlamb/repos/LightGBM/src/boosting/gbdt.cpp should add these lines:
#include <__fwd/fstream.h>                  // for ifstream
#include <cmath>                            // for fabs
#include <cstring>                          // for memset
#include <fstream>                          // for operator<<, basic_ostream
#include <functional>                       // for function
#include <ratio>                            // for milli
#include "LightGBM/cuda/vector_cudahost.h"  // for LGBM_config_, Use_Learner
#include "LightGBM/utils/json11.h"          // for Json
#include "omp.h"                            // for omp_get_thread_num

/Users/jlamb/repos/LightGBM/src/boosting/gbdt.cpp should remove these lines:
- #include <LightGBM/prediction_early_stop.h>  // lines 10-10
- #include <ctime>  // lines 16-16

The full include-list for /Users/jlamb/repos/LightGBM/src/boosting/gbdt.cpp:
#include "gbdt.h"
#include <LightGBM/metric.h>                // for Metric
#include <LightGBM/network.h>               // for Network
#include <LightGBM/objective_function.h>    // for ObjectiveFunction
#include <LightGBM/sample_strategy.h>       // for SampleStrategy
#include <LightGBM/utils/common.h>          // for AlignmentAllocator, Funct...
#include <LightGBM/utils/openmp_wrapper.h>  // for OMP_NUM_THREADS
#include <__fwd/fstream.h>                  // for ifstream
#include <chrono>                           // for duration, operator*, oper...
#include <cmath>                            // for fabs
#include <cstring>                          // for memset
#include <fstream>                          // for operator<<, basic_ostream
#include <functional>                       // for function
#include <queue>                            // for queue
#include <ratio>                            // for milli
#include <sstream>                          // for basic_stringstream
#include "LightGBM/cuda/vector_cudahost.h"  // for LGBM_config_, Use_Learner
#include "LightGBM/utils/json11.h"          // for Json
#include "omp.h"                            // for omp_get_thread_num
---

[ 11%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt_model_text.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/boosting/gbdt_model_text.cpp should add these lines:
#include <__fwd/fstream.h>                  // for ifstream, ofstream
#include <__fwd/sstream.h>                  // for stringstream
#include <stdint.h>                         // for int8_t
#include <algorithm>                        // for min, max, stable_sort
#include <cstring>                          // for size_t, strlen
#include <fstream>                          // for basic_ofstream, basic_ifs...
#include <iomanip>                          // for operator<<, setprecision
#include <iterator>                         // for istreambuf_iterator
#include <limits>                           // for numeric_limits
#include <memory>                           // for unique_ptr, operator!=
#include <unordered_map>                    // for unordered_map
#include <utility>                          // for pair
#include "LightGBM/tree.h"                  // for Tree
#include "LightGBM/utils/file_io.h"         // for VirtualFileWriter
#include "LightGBM/utils/log.h"             // for Log
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX

/Users/jlamb/repos/LightGBM/src/boosting/gbdt_model_text.cpp should remove these lines:
- #include <LightGBM/metric.h>  // lines 6-6

The full include-list for /Users/jlamb/repos/LightGBM/src/boosting/gbdt_model_text.cpp:
#include <LightGBM/config.h>                // for Config, ParseObjectiveAlias
#include <LightGBM/objective_function.h>    // for ObjectiveFunction
#include <LightGBM/utils/array_args.h>      // for ArrayArgs
#include <LightGBM/utils/common.h>          // for Join, Atoi, C_stringstream
#include <__fwd/fstream.h>                  // for ifstream, ofstream
#include <__fwd/sstream.h>                  // for stringstream
#include <stdint.h>                         // for int8_t
#include <algorithm>                        // for min, max, stable_sort
#include <cstring>                          // for size_t, strlen
#include <fstream>                          // for basic_ofstream, basic_ifs...
#include <iomanip>                          // for operator<<, setprecision
#include <iterator>                         // for istreambuf_iterator
#include <limits>                           // for numeric_limits
#include <memory>                           // for unique_ptr, operator!=
#include <sstream>                          // for operator<<, basic_ostream
#include <string>                           // for char_traits, basic_string
#include <unordered_map>                    // for unordered_map
#include <utility>                          // for pair
#include <vector>                           // for vector
#include "LightGBM/tree.h"                  // for Tree
#include "LightGBM/utils/file_io.h"         // for VirtualFileWriter
#include "LightGBM/utils/log.h"             // for Log
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX
#include "gbdt.h"                           // for GBDT
---

[ 13%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt_prediction.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/boosting/gbdt_prediction.cpp should add these lines:
#include <cstring>                           // for memset
#include <memory>                            // for unique_ptr
#include <unordered_map>                     // for unordered_map
#include <vector>                            // for vector
#include "LightGBM/tree.h"                   // for Tree

/Users/jlamb/repos/LightGBM/src/boosting/gbdt_prediction.cpp should remove these lines:
- #include <LightGBM/utils/openmp_wrapper.h>  // lines 7-7

The full include-list for /Users/jlamb/repos/LightGBM/src/boosting/gbdt_prediction.cpp:
#include <LightGBM/objective_function.h>     // for ObjectiveFunction
#include <LightGBM/prediction_early_stop.h>  // for PredictionEarlyStopInstance
#include <cstring>                           // for memset
#include <memory>                            // for unique_ptr
#include <unordered_map>                     // for unordered_map
#include <vector>                            // for vector
#include "LightGBM/tree.h"                   // for Tree
#include "gbdt.h"                            // for GBDT
---

[ 16%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/prediction_early_stop.cpp.o
Warning: include-what-you-use reported diagnostics:

(/Users/jlamb/repos/LightGBM/include/LightGBM/prediction_early_stop.h has correct #includes/fwd-decls)

/Users/jlamb/repos/LightGBM/src/boosting/prediction_early_stop.cpp should add these lines:
#include <stddef.h>              // for size_t

/Users/jlamb/repos/LightGBM/src/boosting/prediction_early_stop.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/boosting/prediction_early_stop.cpp:
#include <LightGBM/prediction_early_stop.h>
#include <LightGBM/utils/log.h>  // for Log
#include <stddef.h>              // for size_t
#include <algorithm>             // for partial_sort
#include <cmath>                 // for fabs
#include <limits>                // for numeric_limits
#include <vector>                // for vector
---

[ 19%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/sample_strategy.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/sample_strategy.h should add these lines:
#include "LightGBM/meta.h"             // for data_size_t, kAlignedSize, sco...
namespace LightGBM { class ObjectiveFunction; }
namespace LightGBM { class TreeLearner; }
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/include/LightGBM/sample_strategy.h should remove these lines:
- #include <LightGBM/config.h>  // lines 13-13
- #include <LightGBM/objective_function.h>  // lines 16-16
- #include <LightGBM/tree_learner.h>  // lines 15-15
- #include <LightGBM/cuda/cuda_utils.hu>  // lines 9-9

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/sample_strategy.h:
#include <LightGBM/dataset.h>          // for Dataset
#include <LightGBM/utils/common.h>     // for AlignmentAllocator
#include <LightGBM/utils/random.h>     // for Random
#include <LightGBM/utils/threading.h>  // for ParallelPartitionRunner
#include <memory>                      // for unique_ptr
#include <vector>                      // for vector
#include "LightGBM/meta.h"             // for data_size_t, kAlignedSize, sco...
namespace LightGBM { class ObjectiveFunction; }
namespace LightGBM { class TreeLearner; }
namespace LightGBM { struct Config; }
---

/Users/jlamb/repos/LightGBM/src/boosting/sample_strategy.cpp should add these lines:
#include <string>             // for basic_string, operator==, string
#include "LightGBM/config.h"  // for Config
namespace LightGBM { class Dataset; }
namespace LightGBM { class ObjectiveFunction; }

/Users/jlamb/repos/LightGBM/src/boosting/sample_strategy.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/boosting/sample_strategy.cpp:
#include <LightGBM/sample_strategy.h>
#include <string>             // for basic_string, operator==, string
#include "LightGBM/config.h"  // for Config
#include "bagging.hpp"        // for BaggingSampleStrategy
#include "goss.hpp"           // for GOSSStrategy
namespace LightGBM { class Dataset; }
namespace LightGBM { class ObjectiveFunction; }
---

[ 22%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/bin.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/bin.h should add these lines:
#include <__fwd/sstream.h>                  // for stringstream
#include <stddef.h>                         // for size_t
#include <stdint.h>                         // for uint32_t, int32_t, int64_t
#include <cmath>                            // for isnan
#include <iomanip>                          // for operator<<, setprecision
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
namespace LightGBM { struct BinaryWriter; }

/Users/jlamb/repos/LightGBM/include/LightGBM/bin.h should remove these lines:
- #include <LightGBM/utils/file_io.h>  // lines 10-10
- #include <functional>  // lines 14-14

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/bin.h:
#include <LightGBM/meta.h>                  // for data_size_t, score_t, com...
#include <LightGBM/utils/common.h>          // for Join
#include <__fwd/sstream.h>                  // for stringstream
#include <stddef.h>                         // for size_t
#include <stdint.h>                         // for uint32_t, int32_t, int64_t
#include <cmath>                            // for isnan
#include <iomanip>                          // for operator<<, setprecision
#include <limits>                           // for numeric_limits
#include <sstream>                          // for basic_ostream, operator<<
#include <string>                           // for char_traits, allocator
#include <unordered_map>                    // for unordered_map
#include <vector>                           // for vector
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
namespace LightGBM { struct BinaryWriter; }
---

/Users/jlamb/repos/LightGBM/src/io/bin.cpp should add these lines:
#include "LightGBM/meta.h"                 // for kZeroThreshold, data_size_t
#include "LightGBM/utils/binary_writer.h"  // for BinaryWriter
#include "LightGBM/utils/log.h"            // for CHECK_LE, Log, CHECK_GT

/Users/jlamb/repos/LightGBM/src/io/bin.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/io/bin.cpp:
#include <LightGBM/bin.h>
#include <LightGBM/utils/array_args.h>     // for ArrayArgs
#include <LightGBM/utils/common.h>         // for AlignmentAllocator, CheckD...
#include <LightGBM/utils/file_io.h>        // for VirtualFileWriter
#include <algorithm>                       // for max, min, stable_sort
#include <cmath>                           // for fabs, isnan, lround
#include <cstdint>                         // for uint8_t, uint32_t, uint16_t
#include <cstring>                         // for memcpy, size_t
#include "LightGBM/meta.h"                 // for kZeroThreshold, data_size_t
#include "LightGBM/utils/binary_writer.h"  // for BinaryWriter
#include "LightGBM/utils/log.h"            // for CHECK_LE, Log, CHECK_GT
#include "dense_bin.hpp"                   // for DenseBin
#include "multi_val_dense_bin.hpp"         // for MultiValDenseBin
#include "multi_val_sparse_bin.hpp"        // for MultiValSparseBin
#include "sparse_bin.hpp"                  // for SparseBin, SparseBinIterator
---

[ 25%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/config.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/config.h should add these lines:
#include <stddef.h>                 // for size_t
#include <stdint.h>                 // for int32_t, int8_t
#include <__hash_table>             // for operator==
#include <utility>                  // for pair

/Users/jlamb/repos/LightGBM/include/LightGBM/config.h should remove these lines:
- #include <LightGBM/meta.h>  // lines 20-20
- #include <memory>  // lines 26-26

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/config.h:
#include <LightGBM/export.h>        // for LIGHTGBM_EXPORT
#include <LightGBM/utils/common.h>  // for AtofAndCheck, AtoiAndCheck, tolower
#include <LightGBM/utils/log.h>     // for Log
#include <stddef.h>                 // for size_t
#include <stdint.h>                 // for int32_t, int8_t
#include <__hash_table>             // for operator==
#include <algorithm>                // for transform
#include <string>                   // for basic_string, operator==, string
#include <unordered_map>            // for unordered_map, __hash_map_iterator
#include <unordered_set>            // for unordered_set
#include <utility>                  // for pair
#include <vector>                   // for vector
---

/Users/jlamb/repos/LightGBM/src/io/config.cpp should add these lines:
#include <__fwd/sstream.h>                  // for stringstream
#include <cmath>                            // for fabs, pow
#include <memory>                           // for allocator
#include <sstream>                          // for operator<<, basic_ostream
#include "LightGBM/meta.h"                  // for kEpsilon, kZeroThreshold

/Users/jlamb/repos/LightGBM/src/io/config.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/io/config.cpp:
#include <LightGBM/config.h>
#include <LightGBM/cuda/vector_cudahost.h>  // for LGBM_Device, LGBM_config_
#include <LightGBM/utils/common.h>          // for tolower, Join, Split, Trim
#include <LightGBM/utils/log.h>             // for Log, LogLevel
#include <LightGBM/utils/random.h>          // for Random
#include <__fwd/sstream.h>                  // for stringstream
#include <cmath>                            // for fabs, pow
#include <limits>                           // for numeric_limits
#include <memory>                           // for allocator
#include <sstream>                          // for operator<<, basic_ostream
#include "LightGBM/meta.h"                  // for kEpsilon, kZeroThreshold
---

[ 27%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/config_auto.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/io/config_auto.cpp should add these lines:
#include <__fwd/sstream.h>          // for stringstream
#include <stdint.h>                 // for int8_t, int32_t
#include <sstream>                  // for operator<<, basic_ostream, basic_...
#include <string>                   // for basic_string, char_traits, hash
#include <unordered_map>            // for unordered_map
#include <unordered_set>            // for unordered_set
#include <vector>                   // for vector
#include "LightGBM/utils/common.h"  // for Join, StringToArray, ArrayCast
#include "LightGBM/utils/log.h"     // for CHECK_GT, CHECK_GE, CHECK_LE, CHE...

/Users/jlamb/repos/LightGBM/src/io/config_auto.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/io/config_auto.cpp:
#include <LightGBM/config.h>        // for Config
#include <__fwd/sstream.h>          // for stringstream
#include <stdint.h>                 // for int8_t, int32_t
#include <sstream>                  // for operator<<, basic_ostream, basic_...
#include <string>                   // for basic_string, char_traits, hash
#include <unordered_map>            // for unordered_map
#include <unordered_set>            // for unordered_set
#include <vector>                   // for vector
#include "LightGBM/utils/common.h"  // for Join, StringToArray, ArrayCast
#include "LightGBM/utils/log.h"     // for CHECK_GT, CHECK_GE, CHECK_LE, CHE...
---

[ 30%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/dataset.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/dataset.h should add these lines:
#include <__mutex/lock_guard.h>             // for lock_guard
#include <__mutex/mutex.h>                  // for mutex
#include <stddef.h>                         // for size_t
#include <stdint.h>                         // for int32_t, uint32_t, int8_t
#include <algorithm>                        // for replace
#include "LightGBM/bin.h"                   // for BinMapper, hist_t, BinIte...
#include "LightGBM/export.h"                // for LIGHTGBM_EXPORT
#include "LightGBM/utils/common.h"          // for CheckAllowedJSON
#include "LightGBM/utils/log.h"             // for Log
namespace LightGBM { class ArrowChunkedArray; }
namespace LightGBM { struct BinaryWriter; }
namespace LightGBM { struct ByteBuffer; }
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/include/LightGBM/dataset.h should remove these lines:
- #include <LightGBM/arrow.h>  // lines 8-8
- #include <LightGBM/config.h>  // lines 9-9
- #include <LightGBM/utils/byte_buffer.h>  // lines 13-13
- #include <LightGBM/utils/random.h>  // lines 15-15
- #include <LightGBM/utils/text_reader.h>  // lines 16-16
- #include <LightGBM/cuda/cuda_column_data.hpp>  // lines 27-27
- #include <LightGBM/cuda/cuda_metadata.hpp>  // lines 28-28
- #include <mutex>  // lines 22-22

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/dataset.h:
#include <LightGBM/feature_group.h>         // for FeatureGroup
#include <LightGBM/meta.h>                  // for data_size_t, label_t, sco...
#include <LightGBM/train_share_states.h>    // for TrainingShareStates
#include <LightGBM/utils/openmp_wrapper.h>  // for OMP_NUM_THREADS
#include <__mutex/lock_guard.h>             // for lock_guard
#include <__mutex/mutex.h>                  // for mutex
#include <stddef.h>                         // for size_t
#include <stdint.h>                         // for int32_t, uint32_t, int8_t
#include <algorithm>                        // for replace
#include <functional>                       // for function
#include <map>                              // for map
#include <memory>                           // for unique_ptr
#include <string>                           // for basic_string, string
#include <unordered_set>                    // for unordered_set
#include <utility>                          // for pair
#include <vector>                           // for vector
#include "LightGBM/bin.h"                   // for BinMapper, hist_t, BinIte...
#include "LightGBM/export.h"                // for LIGHTGBM_EXPORT
#include "LightGBM/utils/common.h"          // for CheckAllowedJSON
#include "LightGBM/utils/log.h"             // for Log
namespace LightGBM { class ArrowChunkedArray; }
namespace LightGBM { class DatasetLoader; }  // lines 33-33
namespace LightGBM { struct BinaryWriter; }
namespace LightGBM { struct ByteBuffer; }
namespace LightGBM { struct Config; }
---

/Users/jlamb/repos/LightGBM/src/io/dataset.cpp should add these lines:
#include <cstring>                          // for size_t, memset, strlen, NULL
#include <ratio>                            // for ratio, milli
#include "LightGBM/config.h"                // for Config
#include "LightGBM/meta.h"                  // for data_size_t, score_t, NO_...
#include "LightGBM/train_share_states.h"    // for TrainingShareStates
#include "LightGBM/utils/binary_writer.h"   // for BinaryWriter
#include "LightGBM/utils/byte_buffer.h"     // for ByteBuffer
#include "LightGBM/utils/file_io.h"         // for VirtualFileWriter
#include "LightGBM/utils/random.h"          // for Random
namespace LightGBM { class ArrowChunkedArray; }

/Users/jlamb/repos/LightGBM/src/io/dataset.cpp should remove these lines:
- #include <limits>  // lines 16-16
- #include <sstream>  // lines 17-17
- #include <unordered_map>  // lines 18-18

The full include-list for /Users/jlamb/repos/LightGBM/src/io/dataset.cpp:
#include <LightGBM/dataset.h>
#include <LightGBM/cuda/vector_cudahost.h>  // for LGBM_Device, LGBM_config_
#include <LightGBM/feature_group.h>         // for FeatureGroup
#include <LightGBM/utils/array_args.h>      // for ArrayArgs
#include <LightGBM/utils/openmp_wrapper.h>  // for OMP_NUM_THREADS, OMP_INIT_EX
#include <LightGBM/utils/threading.h>       // for Threading
#include <chrono>                           // for duration, operator*, oper...
#include <cstdio>                           // for fprintf, size_t, fclose
#include <cstring>                          // for size_t, memset, strlen, NULL
#include <ratio>                            // for ratio, milli
#include "LightGBM/config.h"                // for Config
#include "LightGBM/meta.h"                  // for data_size_t, score_t, NO_...
#include "LightGBM/train_share_states.h"    // for TrainingShareStates
#include "LightGBM/utils/binary_writer.h"   // for BinaryWriter
#include "LightGBM/utils/byte_buffer.h"     // for ByteBuffer
#include "LightGBM/utils/file_io.h"         // for VirtualFileWriter
#include "LightGBM/utils/random.h"          // for Random
namespace LightGBM { class ArrowChunkedArray; }
---

[ 33%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/dataset_loader.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/dataset_loader.h should add these lines:
#include <stddef.h>                 // for size_t
#include <stdint.h>                 // for int32_t, int64_t
#include "LightGBM/export.h"        // for LIGHTGBM_EXPORT
#include "LightGBM/meta.h"          // for data_size_t, PredictFunction
#include "LightGBM/utils/random.h"  // for Random
namespace LightGBM { class BinMapper; }
namespace LightGBM { class Dataset; }
namespace LightGBM { class Metadata; }
namespace LightGBM { class Parser; }
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/include/LightGBM/dataset_loader.h should remove these lines:
- #include <LightGBM/dataset.h>  // lines 8-8

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/dataset_loader.h:
#include <stddef.h>                 // for size_t
#include <stdint.h>                 // for int32_t, int64_t
#include <memory>                   // for unique_ptr
#include <string>                   // for string, basic_string
#include <unordered_set>            // for unordered_set
#include <vector>                   // for vector
#include "LightGBM/export.h"        // for LIGHTGBM_EXPORT
#include "LightGBM/meta.h"          // for data_size_t, PredictFunction
#include "LightGBM/utils/random.h"  // for Random
namespace LightGBM { class BinMapper; }
namespace LightGBM { class Dataset; }
namespace LightGBM { class Metadata; }
namespace LightGBM { class Parser; }
namespace LightGBM { struct Config; }
---

/Users/jlamb/repos/LightGBM/src/io/dataset_loader.cpp should add these lines:
#include <__fwd/fstream.h>                  // for ifstream
#include <__fwd/sstream.h>                  // for stringstream
#include <algorithm>                        // for min_element, min, equal, max
#include <cmath>                            // for fabs, isnan
#include <cstring>                          // for size_t, strlen
#include <functional>                       // for function
#include <ratio>                            // for ratio, milli
#include <sstream>                          // for basic_stringstream
#include <unordered_map>                    // for unordered_map
#include <utility>                          // for pair, move
#include "LightGBM/bin.h"                   // for BinMapper, BinType
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset, Parser, Metadata
#include "LightGBM/feature_group.h"         // for FeatureGroup
#include "LightGBM/utils/common.h"          // for Split, AtoiAndCheck, Star...
#include "LightGBM/utils/file_io.h"         // for VirtualFileReader, Virtua...
#include "LightGBM/utils/text_reader.h"     // for TextReader
#include "LightGBM/utils/threading.h"       // for Threading
#include "omp.h"                            // for omp_get_thread_num

/Users/jlamb/repos/LightGBM/src/io/dataset_loader.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/io/dataset_loader.cpp:
#include <LightGBM/dataset_loader.h>
#include <LightGBM/network.h>               // for Network
#include <LightGBM/utils/array_args.h>      // for ArrayArgs
#include <LightGBM/utils/json11.h>          // for Json
#include <LightGBM/utils/log.h>             // for Log, CHECK, CHECK_EQ, CHE...
#include <LightGBM/utils/openmp_wrapper.h>  // for OMP_NUM_THREADS, OMP_INIT_EX
#include <__fwd/fstream.h>                  // for ifstream
#include <__fwd/sstream.h>                  // for stringstream
#include <algorithm>                        // for min_element, min, equal, max
#include <chrono>                           // for duration, operator*, oper...
#include <cmath>                            // for fabs, isnan
#include <cstring>                          // for size_t, strlen
#include <fstream>                          // for basic_ostream, operator<<
#include <functional>                       // for function
#include <ratio>                            // for ratio, milli
#include <sstream>                          // for basic_stringstream
#include <unordered_map>                    // for unordered_map
#include <utility>                          // for pair, move
#include "LightGBM/bin.h"                   // for BinMapper, BinType
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset, Parser, Metadata
#include "LightGBM/feature_group.h"         // for FeatureGroup
#include "LightGBM/utils/common.h"          // for Split, AtoiAndCheck, Star...
#include "LightGBM/utils/file_io.h"         // for VirtualFileReader, Virtua...
#include "LightGBM/utils/text_reader.h"     // for TextReader
#include "LightGBM/utils/threading.h"       // for Threading
#include "omp.h"                            // for omp_get_thread_num
---

[ 36%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/file_io.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/utils/file_io.h should add these lines:

/Users/jlamb/repos/LightGBM/include/LightGBM/utils/file_io.h should remove these lines:
- #include <cstdlib>  // lines 12-12
- #include <cstring>  // lines 13-13
- #include <iostream>  // lines 14-14
- #include <vector>  // lines 16-16

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/utils/file_io.h:
#include <LightGBM/utils/binary_writer.h>  // for BinaryWriter
#include <cstdio>                          // for size_t
#include <memory>                          // for unique_ptr
#include <string>                          // for string
---

/Users/jlamb/repos/LightGBM/src/io/file_io.cpp should add these lines:

/Users/jlamb/repos/LightGBM/src/io/file_io.cpp should remove these lines:
- #include <LightGBM/utils/log.h>  // lines 8-8
- #include <algorithm>  // lines 10-10
- #include <sstream>  // lines 11-11
- #include <unordered_map>  // lines 12-12

The full include-list for /Users/jlamb/repos/LightGBM/src/io/file_io.cpp:
#include <LightGBM/utils/file_io.h>
---

[ 38%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/json11.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/utils/json11.h should add these lines:
#include <cstddef>           // for nullptr_t, size_t
#include <type_traits>       // for is_constructible, enable_if

/Users/jlamb/repos/LightGBM/include/LightGBM/utils/json11.h should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/utils/json11.h:
#include <cstddef>           // for nullptr_t, size_t
#include <initializer_list>  // for initializer_list
#include <map>               // for map
#include <memory>            // for shared_ptr
#include <string>            // for basic_string, string
#include <type_traits>       // for is_constructible, enable_if
#include <utility>           // for declval, pair
#include <vector>            // for vector
namespace json11_internal_lightgbm { class JsonValue; }  // lines 67-67
---

/Users/jlamb/repos/LightGBM/src/io/json11.cpp should add these lines:
#include <stdint.h>              // for uint8_t, int64_t

/Users/jlamb/repos/LightGBM/src/io/json11.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/io/json11.cpp:
#include <LightGBM/utils/json11.h>
#include <LightGBM/utils/log.h>  // for CHECK_NE
#include <stdint.h>              // for uint8_t, int64_t
#include <cmath>                 // for isfinite
#include <cstdio>                // for size_t, snprintf
#include <cstdlib>               // for atoi, strtod, strtol
#include <limits>                // for numeric_limits
---

[ 41%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/metadata.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/io/metadata.cpp should add these lines:
#include <__mutex/lock_guard.h>             // for lock_guard
#include <__mutex/mutex.h>                  // for mutex
#include <stdint.h>                         // for int32_t, int64_t
#include <cstring>                          // for size_t, memcpy
#include <unordered_map>                    // for unordered_map
#include "LightGBM/arrow.h"                 // for ArrowChunkedArray
#include "LightGBM/arrow.tpp"               // for operator-, ArrowChunkedAr...
#include "LightGBM/meta.h"                  // for data_size_t, label_t
#include "LightGBM/utils/binary_writer.h"   // for BinaryWriter
#include "LightGBM/utils/file_io.h"         // for VirtualFileWriter
#include "LightGBM/utils/log.h"             // for Log
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
#include "LightGBM/utils/text_reader.h"     // for TextReader

/Users/jlamb/repos/LightGBM/src/io/metadata.cpp should remove these lines:
- #include <set>  // lines 8-8

The full include-list for /Users/jlamb/repos/LightGBM/src/io/metadata.cpp:
#include <LightGBM/dataset.h>               // for Metadata
#include <LightGBM/utils/common.h>          // for AvoidInf, Atof, Split, Atoi
#include <__mutex/lock_guard.h>             // for lock_guard
#include <__mutex/mutex.h>                  // for mutex
#include <stdint.h>                         // for int32_t, int64_t
#include <cstring>                          // for size_t, memcpy
#include <string>                           // for basic_string, string, hash
#include <unordered_map>                    // for unordered_map
#include <vector>                           // for vector
#include "LightGBM/arrow.h"                 // for ArrowChunkedArray
#include "LightGBM/arrow.tpp"               // for operator-, ArrowChunkedAr...
#include "LightGBM/meta.h"                  // for data_size_t, label_t
#include "LightGBM/utils/binary_writer.h"   // for BinaryWriter
#include "LightGBM/utils/file_io.h"         // for VirtualFileWriter
#include "LightGBM/utils/log.h"             // for Log
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
#include "LightGBM/utils/text_reader.h"     // for TextReader
---

[ 44%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/parser.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/io/parser.hpp should add these lines:
#include <cmath>                    // for fabs, isnan
#include "LightGBM/meta.h"          // for kZeroThreshold

/Users/jlamb/repos/LightGBM/src/io/parser.hpp should remove these lines:
- #include <unordered_map>  // lines 12-12

The full include-list for /Users/jlamb/repos/LightGBM/src/io/parser.hpp:
#include <LightGBM/dataset.h>       // for Parser
#include <LightGBM/utils/common.h>  // for SkipSpaceAndTab, Atoi, Atof
#include <LightGBM/utils/log.h>     // for Log
#include <cmath>                    // for fabs, isnan
#include <utility>                  // for pair
#include <vector>                   // for vector
#include "LightGBM/meta.h"          // for kZeroThreshold
---

/Users/jlamb/repos/LightGBM/src/io/parser.cpp should add these lines:
#include <__fwd/sstream.h>               // for stringstream
#include <stddef.h>                      // for size_t
#include <sstream>                       // for basic_stringstream, getline
#include "LightGBM/dataset.h"            // for Parser, ParserFactory
#include "LightGBM/utils/common.h"       // for Trim, GetFromParserConfig
#include "LightGBM/utils/file_io.h"      // for VirtualFileReader
#include "LightGBM/utils/log.h"          // for Log, CHECK_GT
#include "LightGBM/utils/text_reader.h"  // for TextReader

/Users/jlamb/repos/LightGBM/src/io/parser.cpp should remove these lines:
- #include <algorithm>  // lines 9-9

The full include-list for /Users/jlamb/repos/LightGBM/src/io/parser.cpp:
#include "parser.hpp"
#include <__fwd/sstream.h>               // for stringstream
#include <stddef.h>                      // for size_t
#include <functional>                    // for function
#include <map>                           // for map, __map_const_iterator
#include <memory>                        // for unique_ptr
#include <sstream>                       // for basic_stringstream, getline
#include <string>                        // for basic_string, string, to_string
#include "LightGBM/dataset.h"            // for Parser, ParserFactory
#include "LightGBM/utils/common.h"       // for Trim, GetFromParserConfig
#include "LightGBM/utils/file_io.h"      // for VirtualFileReader
#include "LightGBM/utils/log.h"          // for Log, CHECK_GT
#include "LightGBM/utils/text_reader.h"  // for TextReader
---

[ 47%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/train_share_states.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/train_share_states.h should add these lines:
#include <stdint.h>                         // for int8_t, int16_t, int32_t
#include <cstring>                          // for memset, size_t
#include <string>                           // for basic_string
#include "LightGBM/utils/common.h"          // for AlignmentAllocator, Timer
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_INIT_EX, OMP_LOOP_EX_...
namespace LightGBM { class FeatureGroup; }

/Users/jlamb/repos/LightGBM/include/LightGBM/train_share_states.h should remove these lines:
- #include <LightGBM/feature_group.h>  // lines 9-9

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/train_share_states.h:
#include <LightGBM/bin.h>                   // for hist_t, MultiValBin
#include <LightGBM/meta.h>                  // for data_size_t, kAlignedSize
#include <LightGBM/utils/threading.h>       // for Threading
#include <stdint.h>                         // for int8_t, int16_t, int32_t
#include <algorithm>                        // for min
#include <cstring>                          // for memset, size_t
#include <memory>                           // for unique_ptr, operator!=
#include <string>                           // for basic_string
#include <vector>                           // for vector
#include "LightGBM/utils/common.h"          // for AlignmentAllocator, Timer
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_INIT_EX, OMP_LOOP_EX_...
namespace LightGBM { class FeatureGroup; }
---

/Users/jlamb/repos/LightGBM/src/io/train_share_states.cpp should add these lines:
#include "LightGBM/bin.h"              // for BinMapper, MultiValBin, hist_t
#include "LightGBM/feature_group.h"    // for FeatureGroup
#include "LightGBM/meta.h"             // for kAlignedSize, data_size_t, kZe...
#include "LightGBM/utils/log.h"        // for CHECK, CHECK_EQ
#include "LightGBM/utils/threading.h"  // for Threading

/Users/jlamb/repos/LightGBM/src/io/train_share_states.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/io/train_share_states.cpp:
#include <LightGBM/train_share_states.h>
#include "LightGBM/bin.h"              // for BinMapper, MultiValBin, hist_t
#include "LightGBM/feature_group.h"    // for FeatureGroup
#include "LightGBM/meta.h"             // for kAlignedSize, data_size_t, kZe...
#include "LightGBM/utils/log.h"        // for CHECK, CHECK_EQ
#include "LightGBM/utils/threading.h"  // for Threading
---

[ 50%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/tree.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/tree.h should add these lines:
#include <stddef.h>                         // for size_t
#include <stdint.h>                         // for uint32_t, int8_t, uint8_t
#include <cmath>                            // for isnan
#include <utility>                          // for pair
#include "LightGBM/bin.h"                   // for MissingType
#include "LightGBM/utils/common.h"          // for FindInBitset
#include "LightGBM/utils/log.h"             // for CHECK_GE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
namespace LightGBM { class Dataset; }

/Users/jlamb/repos/LightGBM/include/LightGBM/tree.h should remove these lines:
- #include <LightGBM/dataset.h>  // lines 8-8
- #include <map>  // lines 12-12
- #include <memory>  // lines 13-13

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/tree.h:
#include <LightGBM/meta.h>                  // for data_size_t, kZeroThreshold
#include <stddef.h>                         // for size_t
#include <stdint.h>                         // for uint32_t, int8_t, uint8_t
#include <cmath>                            // for isnan
#include <string>                           // for string
#include <unordered_map>                    // for unordered_map, __hash_map...
#include <utility>                          // for pair
#include <vector>                           // for vector
#include "LightGBM/bin.h"                   // for MissingType
#include "LightGBM/utils/common.h"          // for FindInBitset
#include "LightGBM/utils/log.h"             // for CHECK_GE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
namespace LightGBM { class Dataset; }
---

/Users/jlamb/repos/LightGBM/src/io/tree.cpp should add these lines:
#include <__fwd/sstream.h>             // for stringstream
#include <algorithm>                   // for copy
#include <limits>                      // for numeric_limits
#include <memory>                      // for unique_ptr, allocator
#include "LightGBM/meta.h"             // for data_size_t

/Users/jlamb/repos/LightGBM/src/io/tree.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/io/tree.cpp:
#include <LightGBM/tree.h>
#include <LightGBM/dataset.h>          // for Dataset
#include <LightGBM/utils/common.h>     // for ArrayToString, StringToArrayFast
#include <LightGBM/utils/threading.h>  // for Threading
#include <__fwd/sstream.h>             // for stringstream
#include <algorithm>                   // for copy
#include <functional>                  // for function
#include <iomanip>                     // for operator<<, setprecision
#include <limits>                      // for numeric_limits
#include <memory>                      // for unique_ptr, allocator
#include <sstream>                     // for operator<<, basic_ostream, bas...
#include "LightGBM/meta.h"             // for data_size_t
---

[ 52%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/metric/dcg_calculator.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/metric/dcg_calculator.cpp should add these lines:
#include <stddef.h>              // for size_t
#include "LightGBM/dataset.h"    // for Metadata
#include "LightGBM/meta.h"       // for data_size_t, label_t, kEpsilon

/Users/jlamb/repos/LightGBM/src/metric/dcg_calculator.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/metric/dcg_calculator.cpp:
#include <LightGBM/metric.h>     // for DCGCalculator
#include <LightGBM/utils/log.h>  // for Log, CHECK_GT
#include <stddef.h>              // for size_t
#include <algorithm>             // for stable_sort
#include <cmath>                 // for fabs, log2
#include <vector>                // for vector
#include "LightGBM/dataset.h"    // for Metadata
#include "LightGBM/meta.h"       // for data_size_t, label_t, kEpsilon
---

[ 55%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/metric/metric.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/metric.h should add these lines:
#include "LightGBM/export.h"  // for LIGHTGBM_EXPORT
namespace LightGBM { class Metadata; }
namespace LightGBM { class ObjectiveFunction; }
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/include/LightGBM/metric.h should remove these lines:
- #include <LightGBM/config.h>  // lines 8-8
- #include <LightGBM/dataset.h>  // lines 9-9
- #include <LightGBM/objective_function.h>  // lines 11-11
- #include <LightGBM/utils/common.h>  // lines 13-13
- #include <LightGBM/utils/log.h>  // lines 12-12

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/metric.h:
#include <LightGBM/meta.h>    // for data_size_t, label_t
#include <string>             // for string
#include <vector>             // for vector
#include "LightGBM/export.h"  // for LIGHTGBM_EXPORT
namespace LightGBM { class Metadata; }
namespace LightGBM { class ObjectiveFunction; }
namespace LightGBM { struct Config; }
---

/Users/jlamb/repos/LightGBM/src/metric/metric.cpp should add these lines:
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/src/metric/metric.cpp should remove these lines:
- #include "cuda/cuda_binary_metric.hpp"  // lines 14-14
- #include "cuda/cuda_regression_metric.hpp"  // lines 15-15

The full include-list for /Users/jlamb/repos/LightGBM/src/metric/metric.cpp:
#include <LightGBM/metric.h>
#include "binary_metric.hpp"      // for AUCMetric, AveragePrecisionMetric
#include "map_metric.hpp"         // for MapMetric
#include "multiclass_metric.hpp"  // for AucMuMetric, MultiErrorMetric, Mult...
#include "rank_metric.hpp"        // for NDCGMetric
#include "regression_metric.hpp"  // for FairLossMetric, GammaDevianceMetric
#include "xentropy_metric.hpp"    // for CrossEntropyLambdaMetric, CrossEntr...
namespace LightGBM { struct Config; }
---

[ 58%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/network/linker_topo.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/network/linker_topo.cpp should add these lines:
#include <stddef.h>            // for size_t

/Users/jlamb/repos/LightGBM/src/network/linker_topo.cpp should remove these lines:
- #include <LightGBM/utils/common.h>  // lines 6-6
- #include <LightGBM/utils/log.h>  // lines 7-7
- #include <string>  // lines 9-9
- #include <unordered_map>  // lines 10-10

The full include-list for /Users/jlamb/repos/LightGBM/src/network/linker_topo.cpp:
#include <LightGBM/network.h>  // for RecursiveHalvingMap, RecursiveHalvingN...
#include <stddef.h>            // for size_t
#include <vector>              // for vector
---

[ 61%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/network/linkers_mpi.cpp.o
[ 63%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/network/linkers_socket.cpp.o
Warning: include-what-you-use reported diagnostics:
/Users/jlamb/repos/LightGBM/src/network/linkers_socket.cpp:174:7: warning: variable 'need_connect_cnt' set but not used [-Wunused-but-set-variable]
  174 |   int need_connect_cnt = 0;
      |       ^

/Users/jlamb/repos/LightGBM/src/network/linkers_socket.cpp should add these lines:
#include <stdlib.h>                      // for atoi
#include <memory>                        // for unique_ptr, default_delete
#include <ratio>                         // for ratio, milli
#include <utility>                       // for pair
#include "LightGBM/network.h"            // for BruckMap, RecursiveHalvingMap
#include "LightGBM/utils/log.h"          // for Log
#include "socket_wrapper.hpp"            // for TcpSocket

/Users/jlamb/repos/LightGBM/src/network/linkers_socket.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/network/linkers_socket.cpp:
#include <LightGBM/config.h>             // for Config
#include <LightGBM/utils/common.h>       // for Split, Trim, Atoi
#include <LightGBM/utils/text_reader.h>  // for TextReader
#include <stdlib.h>                      // for atoi
#include <algorithm>                     // for max
#include <chrono>                        // for duration, operator*, millise...
#include <cstring>                       // for size_t
#include <memory>                        // for unique_ptr, default_delete
#include <ratio>                         // for ratio, milli
#include <string>                        // for basic_string, string
#include <thread>                        // for sleep_for, thread
#include <unordered_map>                 // for unordered_map, __hash_map_it...
#include <unordered_set>                 // for unordered_set
#include <utility>                       // for pair
#include <vector>                        // for vector
#include "LightGBM/network.h"            // for BruckMap, RecursiveHalvingMap
#include "LightGBM/utils/log.h"          // for Log
#include "linkers.h"                     // for Linkers
#include "socket_wrapper.hpp"            // for TcpSocket
---

[ 66%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/network/network.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/network.h should add these lines:
#include <cstring>               // for memcpy
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/include/LightGBM/network.h should remove these lines:
- #include <LightGBM/config.h>  // lines 8-8
- #include <functional>  // lines 12-12

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/network.h:
#include <LightGBM/meta.h>       // for comm_size_t, ReduceFunction, Allgath...
#include <LightGBM/utils/log.h>  // for THREAD_LOCAL
#include <cstring>               // for memcpy
#include <memory>                // for unique_ptr
#include <vector>                // for vector
namespace LightGBM { class Linkers; }  // lines 19-19
namespace LightGBM { struct Config; }
---

/Users/jlamb/repos/LightGBM/src/network/network.cpp should add these lines:
#include <algorithm>             // for reverse, min
#include "LightGBM/config.h"     // for Config
#include "LightGBM/meta.h"       // for comm_size_t, ReduceFunction, Allgath...
#include "LightGBM/utils/log.h"  // for THREAD_LOCAL, Log

/Users/jlamb/repos/LightGBM/src/network/network.cpp should remove these lines:
- #include <LightGBM/utils/common.h>  // lines 7-7
- #include <cstdlib>  // lines 9-9

The full include-list for /Users/jlamb/repos/LightGBM/src/network/network.cpp:
#include <LightGBM/network.h>
#include <algorithm>             // for reverse, min
#include <cstring>               // for memcpy
#include "LightGBM/config.h"     // for Config
#include "LightGBM/meta.h"       // for comm_size_t, ReduceFunction, Allgath...
#include "LightGBM/utils/log.h"  // for THREAD_LOCAL, Log
#include "linkers.h"             // for Linkers
---

[ 69%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/objective/objective_function.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/objective_function.h should add these lines:
#include "LightGBM/export.h"  // for LIGHTGBM_EXPORT
namespace LightGBM { class Metadata; }
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/include/LightGBM/objective_function.h should remove these lines:
- #include <LightGBM/config.h>  // lines 8-8
- #include <LightGBM/dataset.h>  // lines 9-9

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/objective_function.h:
#include <LightGBM/meta.h>    // for data_size_t, score_t, label_t
#include <functional>         // for function
#include <string>             // for string
#include "LightGBM/export.h"  // for LIGHTGBM_EXPORT
namespace LightGBM { class Metadata; }
namespace LightGBM { struct Config; }
---

/Users/jlamb/repos/LightGBM/src/objective/objective_function.cpp should add these lines:
#include <vector>                    // for vector
#include "LightGBM/utils/common.h"   // for Split
#include "LightGBM/utils/log.h"      // for Log
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/src/objective/objective_function.cpp should remove these lines:
- #include "cuda/cuda_binary_objective.hpp"  // lines 13-13
- #include "cuda/cuda_multiclass_objective.hpp"  // lines 14-14
- #include "cuda/cuda_rank_objective.hpp"  // lines 15-15
- #include "cuda/cuda_regression_objective.hpp"  // lines 16-16

The full include-list for /Users/jlamb/repos/LightGBM/src/objective/objective_function.cpp:
#include <LightGBM/objective_function.h>
#include <vector>                    // for vector
#include "LightGBM/utils/common.h"   // for Split
#include "LightGBM/utils/log.h"      // for Log
#include "binary_objective.hpp"      // for BinaryLogloss
#include "multiclass_objective.hpp"  // for MulticlassOVA, MulticlassSoftmax
#include "rank_objective.hpp"        // for LambdarankNDCG, RankXENDCG
#include "regression_objective.hpp"  // for RegressionFairLoss, RegressionGa...
#include "xentropy_objective.hpp"    // for CrossEntropy, CrossEntropyLambda
namespace LightGBM { struct Config; }
---

[ 72%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/data_parallel_tree_learner.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/treelearner/data_parallel_tree_learner.cpp should add these lines:
#include <stdint.h>                         // for int64_t, uint8_t, int32_t
#include <algorithm>                        // for max, min
#include <memory>                           // for unique_ptr, operator!=
#include <string>                           // for basic_string
#include "LightGBM/bin.h"                   // for BinMapper, hist_t, kHistE...
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for comm_size_t, data_size_t
#include "LightGBM/network.h"               // for Network
#include "LightGBM/train_share_states.h"    // for TrainingShareStates
#include "LightGBM/utils/array_args.h"      // for ArrayArgs
#include "LightGBM/utils/common.h"          // for AlignmentAllocator, Timer
#include "LightGBM/utils/log.h"             // for CHECK_LE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX
#include "col_sampler.hpp"                  // for ColSampler
#include "data_partition.hpp"               // for DataPartition
#include "feature_histogram.hpp"            // for FeatureHistogram
#include "gpu_tree_learner.h"               // for GPUTreeLearner
#include "gradient_discretizer.hpp"         // for GradientDiscretizer
#include "leaf_splits.hpp"                  // for LeafSplits
#include "omp.h"                            // for omp_get_thread_num
#include "serial_tree_learner.h"            // for SerialTreeLearner
#include "split_info.hpp"                   // for SplitInfo
namespace LightGBM { class Tree; }

/Users/jlamb/repos/LightGBM/src/treelearner/data_parallel_tree_learner.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/data_parallel_tree_learner.cpp:
#include <stdint.h>                         // for int64_t, uint8_t, int32_t
#include <algorithm>                        // for max, min
#include <cstring>                          // for memcpy, memset, size_t
#include <memory>                           // for unique_ptr, operator!=
#include <string>                           // for basic_string
#include <tuple>                            // for get, tuple
#include <vector>                           // for vector
#include "LightGBM/bin.h"                   // for BinMapper, hist_t, kHistE...
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for comm_size_t, data_size_t
#include "LightGBM/network.h"               // for Network
#include "LightGBM/train_share_states.h"    // for TrainingShareStates
#include "LightGBM/utils/array_args.h"      // for ArrayArgs
#include "LightGBM/utils/common.h"          // for AlignmentAllocator, Timer
#include "LightGBM/utils/log.h"             // for CHECK_LE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX
#include "col_sampler.hpp"                  // for ColSampler
#include "data_partition.hpp"               // for DataPartition
#include "feature_histogram.hpp"            // for FeatureHistogram
#include "gpu_tree_learner.h"               // for GPUTreeLearner
#include "gradient_discretizer.hpp"         // for GradientDiscretizer
#include "leaf_splits.hpp"                  // for LeafSplits
#include "omp.h"                            // for omp_get_thread_num
#include "parallel_tree_learner.h"          // for DataParallelTreeLearner
#include "serial_tree_learner.h"            // for SerialTreeLearner
#include "split_info.hpp"                   // for SplitInfo
namespace LightGBM { class Tree; }
---

[ 75%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/feature_histogram.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/treelearner/feature_histogram.hpp should add these lines:
#include <stdint.h>                         // for int32_t, int64_t, uint32_t
#include <functional>                       // for function
#include "LightGBM/config.h"                // for Config
#include "LightGBM/meta.h"                  // for data_size_t, kEpsilon
#include "LightGBM/utils/common.h"          // for AlignmentAllocator, RoundInt
#include "LightGBM/utils/log.h"             // for CHECK_LE, Log, CHECK_GE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX
#include "LightGBM/utils/random.h"          // for Random

/Users/jlamb/repos/LightGBM/src/treelearner/feature_histogram.hpp should remove these lines:
- #include <utility>  // lines 17-17

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/feature_histogram.hpp:
#include <LightGBM/bin.h>                   // for MissingType, hist_t, GET_...
#include <LightGBM/dataset.h>               // for Dataset
#include <LightGBM/utils/array_args.h>      // for ArrayArgs
#include <stdint.h>                         // for int32_t, int64_t, uint32_t
#include <algorithm>                        // for fill, max
#include <cmath>                            // for fabs, isnan, NAN
#include <cstring>                          // for memcpy
#include <functional>                       // for function
#include <memory>                           // for unique_ptr, swap
#include <vector>                           // for vector, operator!=
#include "LightGBM/config.h"                // for Config
#include "LightGBM/meta.h"                  // for data_size_t, kEpsilon
#include "LightGBM/utils/common.h"          // for AlignmentAllocator, RoundInt
#include "LightGBM/utils/log.h"             // for CHECK_LE, Log, CHECK_GE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX
#include "LightGBM/utils/random.h"          // for Random
#include "monotone_constraints.hpp"         // for BasicConstraint, FeatureC...
#include "split_info.hpp"                   // for SplitInfo
---

/Users/jlamb/repos/LightGBM/src/treelearner/feature_histogram.cpp should add these lines:
#include "LightGBM/bin.h"  // for GET_HESS, GET_GRAD

/Users/jlamb/repos/LightGBM/src/treelearner/feature_histogram.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/feature_histogram.cpp:
#include "feature_histogram.hpp"
#include "LightGBM/bin.h"  // for GET_HESS, GET_GRAD
---

[ 77%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/feature_parallel_tree_learner.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/treelearner/feature_parallel_tree_learner.cpp should add these lines:
#include <stdint.h>                     // for int8_t
#include <memory>                       // for unique_ptr
#include "LightGBM/config.h"            // for Config
#include "LightGBM/dataset.h"           // for Dataset
#include "LightGBM/network.h"           // for Network
#include "LightGBM/utils/array_args.h"  // for ArrayArgs
#include "col_sampler.hpp"              // for ColSampler
#include "gpu_tree_learner.h"           // for GPUTreeLearner
#include "leaf_splits.hpp"              // for LeafSplits
#include "serial_tree_learner.h"        // for SerialTreeLearner
#include "split_info.hpp"               // for SplitInfo
namespace LightGBM { class Tree; }

/Users/jlamb/repos/LightGBM/src/treelearner/feature_parallel_tree_learner.cpp should remove these lines:
- #include <cstring>  // lines 5-5

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/feature_parallel_tree_learner.cpp:
#include <stdint.h>                     // for int8_t
#include <memory>                       // for unique_ptr
#include <vector>                       // for vector
#include "LightGBM/config.h"            // for Config
#include "LightGBM/dataset.h"           // for Dataset
#include "LightGBM/network.h"           // for Network
#include "LightGBM/utils/array_args.h"  // for ArrayArgs
#include "col_sampler.hpp"              // for ColSampler
#include "gpu_tree_learner.h"           // for GPUTreeLearner
#include "leaf_splits.hpp"              // for LeafSplits
#include "parallel_tree_learner.h"      // for FeatureParallelTreeLearner
#include "serial_tree_learner.h"        // for SerialTreeLearner
#include "split_info.hpp"               // for SplitInfo
namespace LightGBM { class Tree; }
---

[ 80%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/gpu_tree_learner.cpp.o
[ 83%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/gradient_discretizer.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/treelearner/gradient_discretizer.hpp should add these lines:
#include <stdint.h>         // for int8_t, int32_t
#include <functional>       // for function
namespace LightGBM { class DataPartition; }
namespace LightGBM { class Dataset; }
namespace LightGBM { class Tree; }
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/src/treelearner/gradient_discretizer.hpp should remove these lines:
- #include <LightGBM/bin.h>  // lines 9-9
- #include <LightGBM/tree.h>  // lines 11-11
- #include <LightGBM/utils/threading.h>  // lines 12-12
- #include "data_partition.hpp"  // lines 17-17
- #include "feature_histogram.hpp"  // lines 18-18

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/gradient_discretizer.hpp:
#include <LightGBM/meta.h>  // for data_size_t, score_t
#include <stdint.h>         // for int8_t, int32_t
#include <functional>       // for function
#include <random>           // for mt19937, uniform_int_distribution
#include <vector>           // for vector
namespace LightGBM { class DataPartition; }
namespace LightGBM { class Dataset; }
namespace LightGBM { class Tree; }
namespace LightGBM { struct Config; }
---

/Users/jlamb/repos/LightGBM/src/treelearner/gradient_discretizer.cpp should add these lines:
#include <cmath>                            // for fabs
#include "LightGBM/bin.h"                   // for BinMapper
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for data_size_t, score_t
#include "LightGBM/tree.h"                  // for Tree
#include "LightGBM/utils/common.h"          // for Timer, global_timer
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
#include "LightGBM/utils/threading.h"       // for Threading
#include "data_partition.hpp"               // for DataPartition
#include "feature_histogram.hpp"            // for FeatureHistogram
#include "omp.h"                            // for omp_get_thread_num

/Users/jlamb/repos/LightGBM/src/treelearner/gradient_discretizer.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/gradient_discretizer.cpp:
#include "gradient_discretizer.hpp"
#include <LightGBM/network.h>               // for Network
#include <algorithm>                        // for min
#include <cmath>                            // for fabs
#include <string>                           // for basic_string, operator==
#include <vector>                           // for vector
#include "LightGBM/bin.h"                   // for BinMapper
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for data_size_t, score_t
#include "LightGBM/tree.h"                  // for Tree
#include "LightGBM/utils/common.h"          // for Timer, global_timer
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
#include "LightGBM/utils/threading.h"       // for Threading
#include "data_partition.hpp"               // for DataPartition
#include "feature_histogram.hpp"            // for FeatureHistogram
#include "omp.h"                            // for omp_get_thread_num
---

[ 86%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/linear_tree_learner.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/treelearner/linear_tree_learner.h should add these lines:
#include <stdint.h>                         // for int8_t
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for score_t
#include "LightGBM/tree.h"                  // for Tree
#include "LightGBM/utils/log.h"             // for CHECK_LE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX
#include "data_partition.hpp"               // for DataPartition
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/src/treelearner/linear_tree_learner.h should remove these lines:
- #include <cstdio>  // lines 9-9
- #include <random>  // lines 11-11

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/linear_tree_learner.h:
#include <stdint.h>                         // for int8_t
#include <cmath>                            // for isnan
#include <memory>                           // for unique_ptr
#include <vector>                           // for vector
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for score_t
#include "LightGBM/tree.h"                  // for Tree
#include "LightGBM/utils/log.h"             // for CHECK_LE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX
#include "data_partition.hpp"               // for DataPartition
#include "serial_tree_learner.h"            // for SerialTreeLearner
namespace LightGBM { struct Config; }
---

/Users/jlamb/repos/LightGBM/src/treelearner/linear_tree_learner.cpp should add these lines:
#include <cstdio>                                             // for size_t
#include <new>                                                // for operato...
#include <type_traits>                                        // for __remov...
#include <utility>                                            // for move, swap
#include "Eigen/src/Core/Assign.h"                            // for MatrixB...
#include "Eigen/src/Core/AssignEvaluator.h"                   // for call_as...
#include "Eigen/src/Core/Block.h"                             // for traits<...
#include "Eigen/src/Core/CoreEvaluators.h"                    // for evaluat...
#include "Eigen/src/Core/CwiseBinaryOp.h"                     // for MatrixB...
#include "Eigen/src/Core/CwiseNullaryOp.h"                    // for traits<...
#include "Eigen/src/Core/CwiseUnaryOp.h"                      // for CwiseUn...
#include "Eigen/src/Core/DenseBase.h"                         // for DenseBa...
#include "Eigen/src/Core/DenseCoeffsBase.h"                   // for inner_s...
#include "Eigen/src/Core/Dot.h"                               // for MatrixB...
#include "Eigen/src/Core/GeneralProduct.h"                    // for product...
#include "Eigen/src/Core/GenericPacketMath.h"                 // for default...
#include "Eigen/src/Core/Inverse.h"                           // for Inverse
#include "Eigen/src/Core/Map.h"                               // for traits<...
#include "Eigen/src/Core/MapBase.h"                           // for MapBase...
#include "Eigen/src/Core/MathFunctions.h"                     // for mini, maxi
#include "Eigen/src/Core/Matrix.h"                            // for Matrix
#include "Eigen/src/Core/MatrixBase.h"                        // for MatrixBase
#include "Eigen/src/Core/NoAlias.h"                           // for MatrixB...
#include "Eigen/src/Core/NumTraits.h"                         // for Generic...
#include "Eigen/src/Core/PartialReduxEvaluator.h"             // for evaluat...
#include "Eigen/src/Core/PermutationMatrix.h"                 // for operator*
#include "Eigen/src/Core/Product.h"                           // for Product
#include "Eigen/src/Core/Redux.h"                             // for DenseBa...
#include "Eigen/src/Core/SelfCwiseBinaryOp.h"                 // for DenseBa...
#include "Eigen/src/Core/SolverBase.h"                        // for SolverB...
#include "Eigen/src/Core/Stride.h"                            // for Stride<...
#include "Eigen/src/Core/Transpose.h"                         // for traits<...
#include "Eigen/src/Core/TriangularMatrix.h"                  // for MatrixB...
#include "Eigen/src/Core/VectorwiseOp.h"                      // for member_...
#include "Eigen/src/Core/arch/SSE/PacketMath.h"               // for unpacke...
#include "Eigen/src/Core/functors/BinaryFunctors.h"           // for functor...
#include "Eigen/src/Core/functors/NullaryFunctors.h"          // for functor...
#include "Eigen/src/Core/functors/UnaryFunctors.h"            // for functor...
#include "Eigen/src/Core/products/GeneralBlockPanelKernel.h"  // for gebp_tr...
#include "Eigen/src/Core/util/BlasUtil.h"                     // for blas_tr...
#include "Eigen/src/Core/util/Constants.h"                    // for Storage...
#include "Eigen/src/Core/util/ForwardDeclarations.h"          // for accesso...
#include "Eigen/src/Core/util/Memory.h"                       // for first_a...
#include "Eigen/src/Core/util/Meta.h"                         // for is_same...
#include "Eigen/src/Core/util/XprHelper.h"                    // for size_at...
#include "Eigen/src/LU/FullPivLU.h"                           // for FullPivLU
#include "LightGBM/bin.h"                                     // for BinType
#include "LightGBM/config.h"                                  // for Config
#include "LightGBM/train_share_states.h"                      // for Trainin...
#include "LightGBM/utils/array_args.h"                        // for ArrayArgs
#include "LightGBM/utils/common.h"                            // for Functio...
#include "monotone_constraints.hpp"                           // for LeafCon...
#include "omp.h"                                              // for omp_get...
#include "split_info.hpp"                                     // for SplitInfo
#include "src/Core/DenseBase.h"                               // for DenseBa...
#include "src/Core/MatrixBase.h"                              // for operator*

/Users/jlamb/repos/LightGBM/src/treelearner/linear_tree_learner.cpp should remove these lines:
- #include <Eigen/Dense>  // lines 7-7

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/linear_tree_learner.cpp:
#include "linear_tree_learner.h"
#include <algorithm>                                          // for fill
#include <cstdio>                                             // for size_t
#include <new>                                                // for operato...
#include <type_traits>                                        // for __remov...
#include <utility>                                            // for move, swap
#include "Eigen/src/Core/Assign.h"                            // for MatrixB...
#include "Eigen/src/Core/AssignEvaluator.h"                   // for call_as...
#include "Eigen/src/Core/Block.h"                             // for traits<...
#include "Eigen/src/Core/CoreEvaluators.h"                    // for evaluat...
#include "Eigen/src/Core/CwiseBinaryOp.h"                     // for MatrixB...
#include "Eigen/src/Core/CwiseNullaryOp.h"                    // for traits<...
#include "Eigen/src/Core/CwiseUnaryOp.h"                      // for CwiseUn...
#include "Eigen/src/Core/DenseBase.h"                         // for DenseBa...
#include "Eigen/src/Core/DenseCoeffsBase.h"                   // for inner_s...
#include "Eigen/src/Core/Dot.h"                               // for MatrixB...
#include "Eigen/src/Core/GeneralProduct.h"                    // for product...
#include "Eigen/src/Core/GenericPacketMath.h"                 // for default...
#include "Eigen/src/Core/Inverse.h"                           // for Inverse
#include "Eigen/src/Core/Map.h"                               // for traits<...
#include "Eigen/src/Core/MapBase.h"                           // for MapBase...
#include "Eigen/src/Core/MathFunctions.h"                     // for mini, maxi
#include "Eigen/src/Core/Matrix.h"                            // for Matrix
#include "Eigen/src/Core/MatrixBase.h"                        // for MatrixBase
#include "Eigen/src/Core/NoAlias.h"                           // for MatrixB...
#include "Eigen/src/Core/NumTraits.h"                         // for Generic...
#include "Eigen/src/Core/PartialReduxEvaluator.h"             // for evaluat...
#include "Eigen/src/Core/PermutationMatrix.h"                 // for operator*
#include "Eigen/src/Core/Product.h"                           // for Product
#include "Eigen/src/Core/Redux.h"                             // for DenseBa...
#include "Eigen/src/Core/SelfCwiseBinaryOp.h"                 // for DenseBa...
#include "Eigen/src/Core/SolverBase.h"                        // for SolverB...
#include "Eigen/src/Core/Stride.h"                            // for Stride<...
#include "Eigen/src/Core/Transpose.h"                         // for traits<...
#include "Eigen/src/Core/TriangularMatrix.h"                  // for MatrixB...
#include "Eigen/src/Core/VectorwiseOp.h"                      // for member_...
#include "Eigen/src/Core/arch/SSE/PacketMath.h"               // for unpacke...
#include "Eigen/src/Core/functors/BinaryFunctors.h"           // for functor...
#include "Eigen/src/Core/functors/NullaryFunctors.h"          // for functor...
#include "Eigen/src/Core/functors/UnaryFunctors.h"            // for functor...
#include "Eigen/src/Core/products/GeneralBlockPanelKernel.h"  // for gebp_tr...
#include "Eigen/src/Core/util/BlasUtil.h"                     // for blas_tr...
#include "Eigen/src/Core/util/Constants.h"                    // for Storage...
#include "Eigen/src/Core/util/ForwardDeclarations.h"          // for accesso...
#include "Eigen/src/Core/util/Memory.h"                       // for first_a...
#include "Eigen/src/Core/util/Meta.h"                         // for is_same...
#include "Eigen/src/Core/util/XprHelper.h"                    // for size_at...
#include "Eigen/src/LU/FullPivLU.h"                           // for FullPivLU
#include "LightGBM/bin.h"                                     // for BinType
#include "LightGBM/config.h"                                  // for Config
#include "LightGBM/train_share_states.h"                      // for Trainin...
#include "LightGBM/utils/array_args.h"                        // for ArrayArgs
#include "LightGBM/utils/common.h"                            // for Functio...
#include "monotone_constraints.hpp"                           // for LeafCon...
#include "omp.h"                                              // for omp_get...
#include "split_info.hpp"                                     // for SplitInfo
#include "src/Core/DenseBase.h"                               // for DenseBa...
#include "src/Core/MatrixBase.h"                              // for operator*
---

[ 88%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/serial_tree_learner.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/treelearner/serial_tree_learner.h should add these lines:
#include <stdint.h>                         // for int8_t, int32_t
#include <functional>                       // for function
#include "LightGBM/meta.h"                  // for score_t, data_size_t, kAl...
#include "LightGBM/train_share_states.h"    // for TrainingShareStates
#include "LightGBM/utils/common.h"          // for AlignmentAllocator
#include "LightGBM/utils/log.h"             // for CHECK_LE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
namespace LightGBM { class Dataset; }
namespace LightGBM { class ObjectiveFunction; }
namespace LightGBM { struct Config; }

/Users/jlamb/repos/LightGBM/src/treelearner/serial_tree_learner.h should remove these lines:
- #include <LightGBM/cuda/vector_cudahost.h>  // lines 11-11
- #include <LightGBM/dataset.h>  // lines 8-8
- #include <LightGBM/utils/array_args.h>  // lines 12-12
- #include <LightGBM/utils/random.h>  // lines 14-14
- #include <cmath>  // lines 17-17
- #include <cstdio>  // lines 18-18
- #include <random>  // lines 20-20
- #include <string>  // lines 16-16

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/serial_tree_learner.h:
#include <LightGBM/tree.h>                  // for Tree
#include <LightGBM/tree_learner.h>          // for TreeLearner
#include <LightGBM/utils/json11.h>          // for Json
#include <stdint.h>                         // for int8_t, int32_t
#include <functional>                       // for function
#include <memory>                           // for unique_ptr
#include <set>                              // for set
#include <vector>                           // for vector
#include "LightGBM/meta.h"                  // for score_t, data_size_t, kAl...
#include "LightGBM/train_share_states.h"    // for TrainingShareStates
#include "LightGBM/utils/common.h"          // for AlignmentAllocator
#include "LightGBM/utils/log.h"             // for CHECK_LE
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS
#include "col_sampler.hpp"                  // for ColSampler
#include "data_partition.hpp"               // for DataPartition
#include "feature_histogram.hpp"            // for HistogramPool
#include "gradient_discretizer.hpp"         // for GradientDiscretizer
#include "leaf_splits.hpp"                  // for LeafSplits
#include "monotone_constraints.hpp"         // for LeafConstraintsBase
#include "split_info.hpp"                   // for SplitInfo
namespace LightGBM { class CostEfficientGradientBoosting; }  // lines 43-43
namespace LightGBM { class Dataset; }
namespace LightGBM { class ObjectiveFunction; }
namespace LightGBM { struct Config; }
---

/Users/jlamb/repos/LightGBM/src/treelearner/serial_tree_learner.cpp should add these lines:
#include <cstdio>                                // for size_t
#include <string>                                // for basic_string, operat...
#include "LightGBM/bin.h"                        // for hist_t, kHistOffset
#include "LightGBM/config.h"                     // for Config
#include "LightGBM/dataset.h"                    // for Dataset
#include "LightGBM/tree.h"                       // for Tree
#include "LightGBM/utils/json11.h"               // for Json
#include "omp.h"                                 // for omp_get_thread_num

/Users/jlamb/repos/LightGBM/src/treelearner/serial_tree_learner.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/serial_tree_learner.cpp:
#include "serial_tree_learner.h"
#include <LightGBM/network.h>                    // for Network
#include <LightGBM/objective_function.h>         // for ObjectiveFunction
#include <LightGBM/utils/array_args.h>           // for ArrayArgs
#include <LightGBM/utils/common.h>               // for AlignmentAllocator
#include <algorithm>                             // for max, min
#include <cstdio>                                // for size_t
#include <queue>                                 // for queue
#include <set>                                   // for set, operator==
#include <string>                                // for basic_string, operat...
#include <unordered_map>                         // for unordered_map, opera...
#include <utility>                               // for pair, make_pair
#include "LightGBM/bin.h"                        // for hist_t, kHistOffset
#include "LightGBM/config.h"                     // for Config
#include "LightGBM/dataset.h"                    // for Dataset
#include "LightGBM/tree.h"                       // for Tree
#include "LightGBM/utils/json11.h"               // for Json
#include "cost_effective_gradient_boosting.hpp"  // for CostEfficientGradien...
#include "omp.h"                                 // for omp_get_thread_num
---

[ 91%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/tree_learner.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/tree_learner.h should add these lines:
#include <functional>       // for function
namespace LightGBM { struct Config; }
namespace json11_internal_lightgbm { class Json; }

/Users/jlamb/repos/LightGBM/include/LightGBM/tree_learner.h should remove these lines:
- #include <LightGBM/config.h>  // lines 8-8
- #include <LightGBM/utils/json11.h>  // lines 10-10

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/tree_learner.h:
#include <LightGBM/meta.h>  // for score_t, data_size_t, label_t
#include <functional>       // for function
#include <string>           // for string
#include <vector>           // for vector
namespace LightGBM { class Dataset; }  // lines 21-21
namespace LightGBM { class ObjectiveFunction; }  // lines 22-22
namespace LightGBM { class Tree; }  // lines 20-20
namespace LightGBM { struct Config; }
namespace json11_internal_lightgbm { class Json; }
---

/Users/jlamb/repos/LightGBM/src/treelearner/tree_learner.cpp should add these lines:
#include "LightGBM/config.h"                      // for Config
#include "LightGBM/utils/log.h"                   // for Log

/Users/jlamb/repos/LightGBM/src/treelearner/tree_learner.cpp should remove these lines:

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/tree_learner.cpp:
#include <LightGBM/tree_learner.h>
#include "LightGBM/config.h"                      // for Config
#include "LightGBM/utils/log.h"                   // for Log
#include "cuda/cuda_single_gpu_tree_learner.hpp"  // for CUDASingleGPUTreeLe...
#include "gpu_tree_learner.h"                     // for GPUTreeLearner
#include "linear_tree_learner.h"                  // for LinearTreeLearner
#include "parallel_tree_learner.h"                // for DataParallelTreeLea...
#include "serial_tree_learner.h"                  // for SerialTreeLearner
---

[ 94%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/voting_parallel_tree_learner.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/src/treelearner/voting_parallel_tree_learner.cpp should add these lines:
#include <stdint.h>                         // for int8_t, uint32_t
#include <algorithm>                        // for max, min, stable_sort
#include <functional>                       // for greater
#include <memory>                           // for unique_ptr, operator!=
#include "LightGBM/bin.h"                   // for BinMapper, hist_t, kHistE...
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for data_size_t, comm_size_t
#include "LightGBM/network.h"               // for Network
#include "LightGBM/train_share_states.h"    // for TrainingShareStates
#include "LightGBM/utils/array_args.h"      // for ArrayArgs
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX
#include "col_sampler.hpp"                  // for ColSampler
#include "data_partition.hpp"               // for DataPartition
#include "feature_histogram.hpp"            // for FeatureHistogram, Feature...
#include "gpu_tree_learner.h"               // for GPUTreeLearner
#include "leaf_splits.hpp"                  // for LeafSplits
#include "omp.h"                            // for omp_get_thread_num
#include "serial_tree_learner.h"            // for SerialTreeLearner
#include "split_info.hpp"                   // for LightSplitInfo, SplitInfo
namespace LightGBM { class Tree; }

/Users/jlamb/repos/LightGBM/src/treelearner/voting_parallel_tree_learner.cpp should remove these lines:
- #include <LightGBM/utils/common.h>  // lines 5-5

The full include-list for /Users/jlamb/repos/LightGBM/src/treelearner/voting_parallel_tree_learner.cpp:
#include <stdint.h>                         // for int8_t, uint32_t
#include <algorithm>                        // for max, min, stable_sort
#include <cstring>                          // for memcpy, size_t, memset
#include <functional>                       // for greater
#include <memory>                           // for unique_ptr, operator!=
#include <tuple>                            // for get, tuple
#include <vector>                           // for vector
#include "LightGBM/bin.h"                   // for BinMapper, hist_t, kHistE...
#include "LightGBM/config.h"                // for Config
#include "LightGBM/dataset.h"               // for Dataset
#include "LightGBM/meta.h"                  // for data_size_t, comm_size_t
#include "LightGBM/network.h"               // for Network
#include "LightGBM/train_share_states.h"    // for TrainingShareStates
#include "LightGBM/utils/array_args.h"      // for ArrayArgs
#include "LightGBM/utils/openmp_wrapper.h"  // for OMP_NUM_THREADS, OMP_INIT_EX
#include "col_sampler.hpp"                  // for ColSampler
#include "data_partition.hpp"               // for DataPartition
#include "feature_histogram.hpp"            // for FeatureHistogram, Feature...
#include "gpu_tree_learner.h"               // for GPUTreeLearner
#include "leaf_splits.hpp"                  // for LeafSplits
#include "omp.h"                            // for omp_get_thread_num
#include "parallel_tree_learner.h"          // for VotingParallelTreeLearner
#include "serial_tree_learner.h"            // for SerialTreeLearner
#include "split_info.hpp"                   // for LightSplitInfo, SplitInfo
namespace LightGBM { class Tree; }
---

[ 97%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/utils/openmp_wrapper.cpp.o
Warning: include-what-you-use reported diagnostics:

/Users/jlamb/repos/LightGBM/include/LightGBM/utils/openmp_wrapper.h should add these lines:
#include <__exception/exception_ptr.h>  // for exception_ptr, operator!=
#include <__exception/operations.h>     // for current_exception, rethrow_ex...
#include <__mutex/mutex.h>              // for mutex
#include <__mutex/unique_lock.h>        // for unique_lock

/Users/jlamb/repos/LightGBM/include/LightGBM/utils/openmp_wrapper.h should remove these lines:
- #include <LightGBM/utils/log.h>  // lines 19-19
- #include <omp.h>  // lines 21-21
- #include <exception>  // lines 23-23
- #include <memory>  // lines 24-24
- #include <mutex>  // lines 25-25
- #include <stdexcept>  // lines 26-26
- #include <vector>  // lines 27-27

The full include-list for /Users/jlamb/repos/LightGBM/include/LightGBM/utils/openmp_wrapper.h:
#include <LightGBM/export.h>            // for LIGHTGBM_EXTERN_C
#include <__exception/exception_ptr.h>  // for exception_ptr, operator!=
#include <__exception/operations.h>     // for current_exception, rethrow_ex...
#include <__mutex/mutex.h>              // for mutex
#include <__mutex/unique_lock.h>        // for unique_lock
---

(/Users/jlamb/repos/LightGBM/src/utils/openmp_wrapper.cpp has correct #includes/fwd-decls)

[ 97%] Built target lightgbm_objs
[100%] Linking CXX shared library ../lib_lightgbm.so
[100%] Built target _lightgbm

[StrikerRUS]

Ah, I thought that cppcheck is useful for such problems. But it seems that I was wrong - no "unused includes" check:

• https://sourceforge.net/p/cppcheck/wiki/ListOfChecks/
• https://sourceforge.net/p/cppcheck/discussion/general/thread/bc8e238be2/
• https://www.reddit.com/r/cprogramming/comments/jy5pn1/comment/gd23y6w/

[StrikerRUS]

I spent about 2 hours trying to make it work in our CI, but unfortunately I wasn't able to make it output anything.
Probably it's the same issue as described here: https://gitlab.kitware.com/cmake/cmake/-/issues/19706#note_626545, but I don't see any compilation warnings in our usual Clang job: https://dev.azure.com/lightgbm-ci/lightgbm-ci/_build/results?buildId=17393&view=logs&j=275189f9-c769-596a-7ef9-49fb48a9ab70&t=3a9e7a4a-04e6-52a0-67ea-6e8f6cfda74f&l=3040

What I've done:

• Added

echo "running include-what-you-use"
# cmake -B build -S . -DCMAKE_C_INCLUDE_WHAT_YOU_USE="include-what-you-use;-Xiwyu;any;-Xiwyu;iwyu;-Xiwyu;args" -DCMAKE_CXX_INCLUDE_WHAT_YOU_USE="include-what-you-use;-Xiwyu;any;-Xiwyu;iwyu;-Xiwyu;args" -G Ninja
# cmake --build build -j4 || true
cmake \
    -B build \
    -S . \
    -DCMAKE_C_COMPILER=clang \
    -DCMAKE_CXX_COMPILER=clang++ \
    -DCMAKE_C_INCLUDE_WHAT_YOU_USE="include-what-you-use;-w;-Xiwyu;--verbose=7" \
    -DCMAKE_CXX_INCLUDE_WHAT_YOU_USE="include-what-you-use;-w;-Xiwyu;--verbose=7"

cmake \
    --build build \
    --target _lightgbm \
    >./logs.txt 2>&1
cat ./logs.txt
echo "done running include-what-you-use"

here

LightGBM/.ci/lint-cpp.sh

Line 4 in 31205fc

• Modified this line to install libomp for lint CI job

if [[ $TASK == "r-package" || $TASK == "lint" ]] && [[ $COMPILER == "clang" ]]; then

LightGBM/.ci/setup.sh

Line 88 in 31205fc

if [[ $TASK == "r-package" ]] && [[ $COMPILER == "clang" ]]; then

• Installed IWYU via conda here

'include-what-you-use>=0.23' \

LightGBM/.ci/test.sh

Line 108 in 31205fc

'cpplint>=1.6.0' \

• Switched to clang in lint job and fetched git submodules for it

COMPILER: 'clang'

LightGBM/.github/workflows/static_analysis.yml

Line 19 in 31205fc

COMPILER: 'gcc'

submodules: true

LightGBM/.github/workflows/static_analysis.yml

Line 40 in 31205fc

submodules: false

And after all these changes and multiple versions of DCMAKE_C_INCLUDE_WHAT_YOU_USE values I got clean build logs without any IWYU outputs

running include-what-you-use
-- The C compiler identification is Clang 14.0.0
-- The CXX compiler identification is Clang 14.0.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/clang - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/clang++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found OpenMP_C: -fopenmp=libomp (found version "5.0")
-- Found OpenMP_CXX: -fopenmp=libomp (found version "5.0")
-- Found OpenMP: TRUE (found version "5.0")
-- Performing Test MM_PREFETCH
-- Performing Test MM_PREFETCH - Success
-- Using _mm_prefetch
-- Performing Test MM_MALLOC
-- Performing Test MM_MALLOC - Success
-- Using _mm_malloc
-- Configuring done (1.7s)
-- Generating done (0.0s)
-- Build files have been written to: /home/runner/work/LightGBM/LightGBM/build
[  2%] Building CXX object CMakeFiles/lightgbm_capi_objs.dir/src/c_api.cpp.o
[  5%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/boosting.cpp.o
[  8%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt_model_text.cpp.o
[ 11%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt_prediction.cpp.o
[ 13%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt.cpp.o
[ 16%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/prediction_early_stop.cpp.o
[ 19%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/sample_strategy.cpp.o
[ 22%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/bin.cpp.o
[ 25%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/config_auto.cpp.o
[ 27%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/config.cpp.o
[ 27%] Built target lightgbm_capi_objs
[ 30%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/dataset_loader.cpp.o
[ 33%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/dataset.cpp.o
[ 36%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/file_io.cpp.o
[ 38%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/json11.cpp.o
[ 41%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/metadata.cpp.o
[ 44%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/parser.cpp.o
[ 47%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/train_share_states.cpp.o
[ 50%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/io/tree.cpp.o
[ 52%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/metric/dcg_calculator.cpp.o
[ 55%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/metric/metric.cpp.o
[ 58%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/network/linker_topo.cpp.o
[ 61%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/network/linkers_mpi.cpp.o
[ 63%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/network/linkers_socket.cpp.o
[ 66%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/network/network.cpp.o
[ 69%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/objective/objective_function.cpp.o
[ 72%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/data_parallel_tree_learner.cpp.o
[ 75%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/feature_histogram.cpp.o
[ 77%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/feature_parallel_tree_learner.cpp.o
[ 80%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/gpu_tree_learner.cpp.o
[ 83%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/gradient_discretizer.cpp.o
[ 86%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/linear_tree_learner.cpp.o
[ 88%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/serial_tree_learner.cpp.o
[ 91%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/tree_learner.cpp.o
[ 94%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/treelearner/voting_parallel_tree_learner.cpp.o
[ 97%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/utils/openmp_wrapper.cpp.o
[ 97%] Built target lightgbm_objs
[100%] Linking CXX shared library /home/runner/work/LightGBM/LightGBM/lib_lightgbm.so
[100%] Built target _lightgbm
done running include-what-you-use

[StrikerRUS]

As an alternative I think we can try clang-tidy's misc-include-cleaner.

Some refs.:

• https://ortogonal.github.io/cmake-clang-tidy/
• https://danielsieger.com/blog/2021/12/21/clang-tidy-cmake.html

Logs:

-- Build files have been written to: /home/runner/work/LightGBM/LightGBM/build
[  2%] Building CXX object CMakeFiles/lightgbm_capi_objs.dir/src/c_api.cpp.o
[  5%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/boosting.cpp.o
[  8%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt_model_text.cpp.o
[ 11%] Building CXX object CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt_prediction.cpp.o
/home/runner/miniforge/envs/test-env/x86_64-conda-linux-gnu/sysroot/usr/include/wchar.h:39:11: error: 'stdarg.h' file not found [clang-diagnostic-error]
   39 | # include <stdarg.h>
      |           ^~~~~~~~~~
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_prediction.cpp:7:1: warning: included header openmp_wrapper.h is not used directly [misc-include-cleaner]
    7 | #include <LightGBM/utils/openmp_wrapper.h>
      | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    8 | 
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_prediction.cpp:16:8: warning: no header providing "std::memset" is directly included [misc-include-cleaner]
    8 | 
    9 | #include "gbdt.h"
   10 | 
   11 | namespace LightGBM {
   12 | 
   13 | void GBDT::PredictRaw(const double* features, double* output, const PredictionEarlyStopInstance* early_stop) const {
   14 |   int early_stop_round_counter = 0;
   15 |   // set zero
   16 |   std::memset(output, 0, sizeof(double) * num_tree_per_iteration_);
      |        ^
2 warnings and 1 error generated.
Error while processing /home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_prediction.cpp.
Found compiler error(s).
/home/runner/miniforge/envs/test-env/x86_64-conda-linux-gnu/sysroot/usr/include/wchar.h:39:11: error: 'stdarg.h' file not found [clang-diagnostic-error]
   39 | # include <stdarg.h>
      |           ^~~~~~~~~~
/home/runner/work/LightGBM/LightGBM/src/boosting/boosting.cpp:13:6: warning: no header providing "std::string" is directly included [misc-include-cleaner]
    6 | 
    7 | #include "dart.hpp"
    8 | #include "gbdt.h"
    9 | #include "rf.hpp"
   10 | 
   11 | namespace LightGBM {
   12 | 
   13 | std::string GetBoostingTypeFromModelFile(const char* filename) {
      |      ^
/home/runner/work/LightGBM/LightGBM/src/boosting/boosting.cpp:20:34: warning: no header providing "std::chrono::steady_clock" is directly included [misc-include-cleaner]
    6 | 
    7 | #include "dart.hpp"
    8 | #include "gbdt.h"
    9 | #include "rf.hpp"
   10 | 
   11 | namespace LightGBM {
   12 | 
   13 | std::string GetBoostingTypeFromModelFile(const char* filename) {
   14 |   TextReader<size_t> model_reader(filename, true);
   15 |   std::string type = model_reader.first_line();
   16 |   return type;
   17 | }
   18 | 
   19 | bool Boosting::LoadFileToBoosting(Boosting* boosting, const char* filename) {
   20 |   auto start_time = std::chrono::steady_clock::now();
      |                                  ^
/home/runner/work/LightGBM/LightGBM/src/boosting/boosting.cpp:29:16: warning: no header providing "std::chrono::duration" is directly included [misc-include-cleaner]
   29 |   std::chrono::duration<double, std::milli> delta = (std::chrono::steady_clock::now() - start_time);
      |                ^
/home/runner/work/LightGBM/LightGBM/src/boosting/boosting.cpp:29:38: warning: no header providing "std::milli" is directly included [misc-include-cleaner]
    6 |   std::chrono::duration<double, std::milli> delta = (std::chrono::steady_clock::now() - start_time);
      |                                      ^
/home/runner/work/LightGBM/LightGBM/src/boosting/boosting.cpp:30:3: warning: no header providing "LightGBM::Log" is directly included [misc-include-cleaner]
    7 |   Log::Debug("Time for loading model: %f seconds", 1e-3*delta);
      |   ^
5 warnings and 1 error generated.
Error while processing /home/runner/work/LightGBM/LightGBM/src/boosting/boosting.cpp.
Found compiler error(s).
gmake[3]: *** [CMakeFiles/lightgbm_objs.dir/build.make:104: CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt_prediction.cpp.o] Error 1
gmake[3]: *** Waiting for unfinished jobs....
gmake[3]: *** [CMakeFiles/lightgbm_objs.dir/build.make:76: CMakeFiles/lightgbm_objs.dir/src/boosting/boosting.cpp.o] Error 1
/home/runner/miniforge/envs/test-env/x86_64-conda-linux-gnu/sysroot/usr/include/wchar.h:39:11: error: 'stdarg.h' file not found [clang-diagnostic-error]
   39 | # include <stdarg.h>
      |           ^~~~~~~~~~
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:5:1: warning: included header config.h is not used directly [misc-include-cleaner]
    5 | #include <LightGBM/config.h>
      | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
    6 | #include <LightGBM/metric.h>
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:6:1: warning: included header metric.h is not used directly [misc-include-cleaner]
    6 | #include <LightGBM/metric.h>
      | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
    7 | #include <LightGBM/objective_function.h>
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:8:1: warning: included header array_args.h is not used directly [misc-include-cleaner]
    8 | #include <LightGBM/utils/array_args.h>
      | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    9 | #include <LightGBM/utils/common.h>
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:13:1: warning: included header vector is not used directly [misc-include-cleaner]
   13 | #include <vector>
      | ^~~~~~~~~~~~~~~~~
   14 | 
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:56:28: warning: no header providing "std::setprecision" is directly included [misc-include-cleaner]
   11 |       json_str_buf << std::setprecision(std::numeric_limits<double>::digits10 + 2);
      |                            ^
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:56:46: warning: no header providing "std::numeric_limits" is directly included [misc-include-cleaner]
   11 |       json_str_buf << std::setprecision(std::numeric_limits<double>::digits10 + 2);
      |                                              ^
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:288:8: warning: no header providing "std::ofstream" is directly included [misc-include-cleaner]
   11 |   std::ofstream output_file;
      |        ^
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:289:8: warning: no header providing "std::ifstream" is directly included [misc-include-cleaner]
  289 |   std::ifstream ifs(filename);
      |        ^
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:357:40: warning: no header providing "OMP_NUM_THREADS" is directly included [misc-include-cleaner]
   15 |   #pragma omp parallel for num_threads(OMP_NUM_THREADS()) schedule(static)
      |                                        ^
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:414:5: warning: no header providing "LightGBM::Log" is directly included [misc-include-cleaner]
   15 |     Log::Fatal("Model file %s is not available for writes", filename);
      |     ^
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:554:5: warning: no header providing "OMP_INIT_EX" is directly included [misc-include-cleaner]
  554 |     OMP_INIT_EX();
      |     ^
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:557:7: warning: no header providing "OMP_LOOP_EX_BEGIN" is directly included [misc-include-cleaner]
  557 |       OMP_LOOP_EX_BEGIN();
      |       ^
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:569:7: warning: no header providing "OMP_LOOP_EX_END" is directly included [misc-include-cleaner]
  569 |       OMP_LOOP_EX_END();
      |       ^
/home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp:571:5: warning: no header providing "OMP_THROW_EX" is directly included [misc-include-cleaner]
  571 |     OMP_THROW_EX();
      |     ^
14 warnings and 1 error generated.
Error while processing /home/runner/work/LightGBM/LightGBM/src/boosting/gbdt_model_text.cpp.
Found compiler error(s).
gmake[3]: *** [CMakeFiles/lightgbm_objs.dir/build.make:90: CMakeFiles/lightgbm_objs.dir/src/boosting/gbdt_model_text.cpp.o] Error 1
gmake[2]: *** [CMakeFiles/Makefile2:89: CMakeFiles/lightgbm_objs.dir/all] Error 2
gmake[2]: *** Waiting for unfinished jobs....
/home/runner/miniforge/envs/test-env/x86_64-conda-linux-gnu/sysroot/usr/include/stdlib.h:32:10: error: 'stddef.h' file not found [clang-diagnostic-error]
   32 | #include <stddef.h>
      |          ^~~~~~~~~~
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:12:1: warning: included header metric.h is not used directly [misc-include-cleaner]
   12 | #include <LightGBM/metric.h>
      | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
   13 | #include <LightGBM/network.h>
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:14:1: warning: included header objective_function.h is not used directly [misc-include-cleaner]
   14 | #include <LightGBM/objective_function.h>
      | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   15 | #include <LightGBM/prediction_early_stop.h>
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:15:1: warning: included header prediction_early_stop.h is not used directly [misc-include-cleaner]
   15 | #include <LightGBM/prediction_early_stop.h>
      | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   16 | #include <LightGBM/utils/byte_buffer.h>
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:21:1: warning: included header threading.h is not used directly [misc-include-cleaner]
   21 | #include <LightGBM/utils/threading.h>
      | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   22 | 
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:27:1: warning: included header memory is not used directly [misc-include-cleaner]
   27 | #include <memory>
      | ^~~~~~~~~~~~~~~~~
   28 | #include <mutex>
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:29:1: warning: included header stdexcept is not used directly [misc-include-cleaner]
   29 | #include <stdexcept>
      | ^~~~~~~~~~~~~~~~~~~~
   30 | #include <vector>
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:30:1: warning: included header vector is not used directly [misc-include-cleaner]
   30 | #include <vector>
      | ^~~~~~~~~~~~~~~~~
   31 | 
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:38:47: warning: no header providing "std::exception" is directly included [misc-include-cleaner]
   23 | #include <string>
   24 | #include <cstdio>
   25 | #include <cstdint>
   26 | #include <functional>
   27 | #include <memory>
   28 | #include <mutex>
   29 | #include <stdexcept>
   30 | #include <vector>
   31 | 
   32 | #include "application/predictor.hpp"
   33 | #include <LightGBM/utils/yamc/alternate_shared_mutex.hpp>
   34 | #include <LightGBM/utils/yamc/yamc_shared_lock.hpp>
   35 | 
   36 | namespace LightGBM {
   37 | 
   38 | inline int LGBM_APIHandleException(const std::exception& ex) {
      |                                               ^
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:65:3: warning: no header providing "LightGBM::PredictFunction" is directly included [misc-include-cleaner]
   32 |   PredictFunction predict_function;
      |   ^
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:416:27: warning: no header providing "LightGBM::score_t" is directly included [misc-include-cleaner]
  416 |   bool TrainOneIter(const score_t* gradients, const score_t* hessians) {
      |                           ^
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:895:17: warning: no header providing "LightGBM::AllgatherFunction" is directly included [misc-include-cleaner]
  895 | using LightGBM::AllgatherFunction;
      |                 ^
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:904:17: warning: no header providing "LightGBM::data_size_t" is directly included [misc-include-cleaner]
  904 | using LightGBM::data_size_t;
      |                 ^
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:907:17: warning: no header providing "LightGBM::kZeroThreshold" is directly included [misc-include-cleaner]
  907 | using LightGBM::kZeroThreshold;
      |                 ^
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:912:17: warning: no header providing "LightGBM::ReduceScatterFunction" is directly included [misc-include-cleaner]
  912 | using LightGBM::ReduceScatterFunction;
      |                 ^
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:[940](https://github.com/microsoft/LightGBM/actions/runs/12342122944/job/34441580895#step:3:941):8: warning: no header providing "std::pair" is directly included [misc-include-cleaner]
   30 |   std::pair<int, double> NextNonZero();
      |        ^
/home/runner/work/LightGBM/LightGBM/src/c_api.cpp:963:10: warning: no header providing "std::memcpy" is directly included [misc-include-cleaner]
   23 |     std::memcpy(out_str, aliases.c_str(), *out_len);
      |          ^
16 warnings and 1 error generated.
Error while processing /home/runner/work/LightGBM/LightGBM/src/c_api.cpp.
Found compiler error(s).
gmake[3]: *** [CMakeFiles/lightgbm_capi_objs.dir/build.make:76: CMakeFiles/lightgbm_capi_objs.dir/src/c_api.cpp.o] Error 1
gmake[2]: *** [CMakeFiles/Makefile2:141: CMakeFiles/lightgbm_capi_objs.dir/all] Error 2
gmake[1]: *** [CMakeFiles/Makefile2:175: CMakeFiles/_lightgbm.dir/rule] Error 2
gmake: *** [Makefile:208: _lightgbm] Error 2

[jameslamb]

Thanks for trying this out! Some time, I'll try this again (as I was able to get it working on my mac back in #6284 (comment)

Separately... I'd support trying to get clang-tidy running in this project if you're interested in pursuing that! I think it could be really helpful, and I trust our coverage of compilers / operating systems in CI here to catch any issues caused by its recommendations.

Some RAPIDS projects have recently started adopting that tool and it seems helpful to me, e.g. https://github.com/rapidsai/legate-boost/pull/186/files#diff-4d2a8eefdf2a9783512a35da4dc7676a66404b6f3826a8af9aad038722da6823

[StrikerRUS]

As a next step I'll

Try reverting the change from #6265 and test if clang-tidy would have caught it.

In case of success, I'll prepare a PR with clang-tidy.

[StrikerRUS]

Unfortunately, clang-tidy wasn't able to catch the error from #6265.

[jameslamb]

hmmmm sad! I would still support adding clang-tidy, those other includes warnings you found in #6284 (comment) look relevant and helpful to me.

[StrikerRUS]

I'm afraid clang-tidy may conflict with include-what-you-use...

[jameslamb]

Oh, I see. Ok well I'd still like to pursue one of these options if we can, even if we sometimes have to use // nolint or similar comments to tell one tool to ignore a line that it and another tool disagree with.