This is an automated email from the ASF dual-hosted git repository.
HuaHuaY pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow.git
The following commit(s) were added to refs/heads/main by this push:
new ab171b294d8 GH-51245: [C++][Parquet][FOLLOWUP] Add support for LLVM
23.1 (#51468)
ab171b294d8 is described below
commit ab171b294d8bfde02254ec0d50dc7afcc32a5416
Author: Zehua Zou <[email protected]>
AuthorDate: Wed Sep 23 18:52:15 2026 +0800
GH-51245: [C++][Parquet][FOLLOWUP] Add support for LLVM 23.1 (#51468)
### Rationale for this change
Add support for LLVM 23.1.
My mistake. In the previous PR, I used the `ninja-debug-gandiva` CMake
preset but didn't notice that some Parquet tests weren't being compiled;
consequently, certain compilation errors within the Parquet component went
undetected. This time I used the `ninja-debug` preset. I'm not sure if there
are compilation errors in other presets—I don't use them often, and CI doesn't
check them either—so we'll just fix them if we run into them.
### What changes are included in this PR?
Remove static from some template functions in header files because LLVM 23
added `-Wunused-template` to `-Wall`.
### Are these changes tested?
Yes.
### Are there any user-facing changes?
Yes.
### Was AI used for this PR?
In accordance to the [AI generation
guidelines](https://arrow.apache.org/docs/dev/developers/overview.html#ai-generated-code),
please disclose below whether and how AI was used in this PR.
**PR code and description written by:**
- [x] Human
- [ ] AI
**Reviewed before submission by:**
- [x] Human
- [ ] AI
- [ ] Not reviewed
* GitHub Issue: #51245
Authored-by: Zehua Zou <[email protected]>
Signed-off-by: Zehua Zou <[email protected]>
---
cpp/src/parquet/arrow/test_util.h | 2 +-
cpp/src/parquet/column_reader_test.cc | 32 ----------------
cpp/src/parquet/test_util.h | 69 ++++++++++++++++-------------------
3 files changed, 32 insertions(+), 71 deletions(-)
diff --git a/cpp/src/parquet/arrow/test_util.h
b/cpp/src/parquet/arrow/test_util.h
index 05f6fd24ac0..2c23e7162be 100644
--- a/cpp/src/parquet/arrow/test_util.h
+++ b/cpp/src/parquet/arrow/test_util.h
@@ -141,7 +141,7 @@ template <typename ArrowType>
}
template <int32_t byte_width>
-static void random_decimals(int64_t n, uint32_t seed, int32_t precision,
uint8_t* out) {
+void random_decimals(int64_t n, uint32_t seed, int32_t precision, uint8_t*
out) {
auto gen = ::arrow::random::RandomArrayGenerator(seed);
std::shared_ptr<Array> decimals;
if constexpr (byte_width == 4) {
diff --git a/cpp/src/parquet/column_reader_test.cc
b/cpp/src/parquet/column_reader_test.cc
index 87514d87db6..e042f6f1d4d 100644
--- a/cpp/src/parquet/column_reader_test.cc
+++ b/cpp/src/parquet/column_reader_test.cc
@@ -48,38 +48,6 @@ using testing::ElementsAre;
namespace test {
-template <typename T>
-static inline bool vector_equal_with_def_levels(const std::vector<T>& left,
- const std::vector<int16_t>&
def_levels,
- int16_t max_def_levels,
- int16_t max_rep_levels,
- const std::vector<T>& right) {
- size_t i_left = 0;
- size_t i_right = 0;
- for (size_t i = 0; i < def_levels.size(); i++) {
- if (def_levels[i] == max_def_levels) {
- // Compare
- if (left[i_left] != right[i_right]) {
- std::cerr << "index " << i << " left was " << left[i_left] << " right
was "
- << right[i] << std::endl;
- return false;
- }
- i_left++;
- i_right++;
- } else if (def_levels[i] == (max_def_levels - 1)) {
- // Null entry on the lowest nested level
- i_right++;
- } else if (def_levels[i] < (max_def_levels - 1)) {
- // Null entry on a higher nesting level, only supported for
non-repeating data
- if (max_rep_levels == 0) {
- i_right++;
- }
- }
- }
-
- return true;
-}
-
class TestPrimitiveReader : public ::testing::Test {
public:
void InitReader(const ColumnDescriptor* d) {
diff --git a/cpp/src/parquet/test_util.h b/cpp/src/parquet/test_util.h
index 9271dc290ca..0e1199378b6 100644
--- a/cpp/src/parquet/test_util.h
+++ b/cpp/src/parquet/test_util.h
@@ -76,8 +76,7 @@ std::string get_bad_data_dir();
std::string get_data_file(const std::string& filename, bool is_good = true);
template <typename T>
-static inline void assert_vector_equal(const std::vector<T>& left,
- const std::vector<T>& right) {
+inline void assert_vector_equal(const std::vector<T>& left, const
std::vector<T>& right) {
ASSERT_EQ(left.size(), right.size());
for (size_t i = 0; i < left.size(); ++i) {
@@ -86,7 +85,7 @@ static inline void assert_vector_equal(const std::vector<T>&
left,
}
template <typename T>
-static inline bool vector_equal(const std::vector<T>& left, const
std::vector<T>& right) {
+inline bool vector_equal(const std::vector<T>& left, const std::vector<T>&
right) {
if (left.size() != right.size()) {
return false;
}
@@ -103,7 +102,7 @@ static inline bool vector_equal(const std::vector<T>& left,
const std::vector<T>
}
template <typename T>
-static std::vector<T> slice(const std::vector<T>& values, int start, int end) {
+std::vector<T> slice(const std::vector<T>& values, int start, int end) {
if (end < start) {
return std::vector<T>(0);
}
@@ -176,21 +175,20 @@ std::shared_ptr<Buffer> EncodeValues(Encoding::type
encoding, bool use_dictionar
}
template <typename T>
-static void InitValues(int num_values, uint32_t seed, std::vector<T>& values,
- std::vector<uint8_t>& buffer) {
+void InitValues(int num_values, uint32_t seed, std::vector<T>& values,
+ std::vector<uint8_t>& buffer) {
random_numbers(num_values, seed, std::numeric_limits<T>::min(),
std::numeric_limits<T>::max(), values.data());
}
template <typename T>
-static void InitValues(int num_values, std::vector<T>& values,
- std::vector<uint8_t>& buffer) {
+void InitValues(int num_values, std::vector<T>& values, std::vector<uint8_t>&
buffer) {
InitValues(num_values, 0, values, buffer);
}
template <typename T>
-static void InitDictValues(int num_values, int num_dicts, std::vector<T>&
values,
- std::vector<uint8_t>& buffer) {
+void InitDictValues(int num_values, int num_dicts, std::vector<T>& values,
+ std::vector<uint8_t>& buffer) {
int repeat_factor = num_values / num_dicts;
InitValues<T>(num_dicts, values, buffer);
// add some repeated values
@@ -347,7 +345,7 @@ inline void
DataPageBuilder<BooleanType>::AppendValues(const ColumnDescriptor* d
}
template <typename Type>
-static std::shared_ptr<DataPageV1> MakeDataPage(
+std::shared_ptr<DataPageV1> MakeDataPage(
const ColumnDescriptor* d, const std::vector<typename Type::c_type>&
values,
int num_vals, Encoding::type encoding, const uint8_t* indices, int
indices_size,
const std::vector<int16_t>& def_levels, int16_t max_def_level,
@@ -441,7 +439,7 @@ inline std::shared_ptr<Buffer>
DictionaryPageBuilder<BooleanType>::AppendValues(
}
template <typename Type>
-inline static std::shared_ptr<DictionaryPage> MakeDictPage(
+inline std::shared_ptr<DictionaryPage> MakeDictPage(
const ColumnDescriptor* d, const std::vector<typename Type::c_type>&
values,
const std::vector<int>& values_per_page, Encoding::type encoding,
std::vector<std::shared_ptr<Buffer>>& rle_indices) {
@@ -463,15 +461,13 @@ inline static std::shared_ptr<DictionaryPage>
MakeDictPage(
// Given def/rep levels and values create multiple dict pages
template <typename Type>
-inline static void PaginateDict(const ColumnDescriptor* d,
- const std::vector<typename Type::c_type>&
values,
- const std::vector<int16_t>& def_levels,
- int16_t max_def_level,
- const std::vector<int16_t>& rep_levels,
- int16_t max_rep_level, int num_levels_per_page,
- const std::vector<int>& values_per_page,
- std::vector<std::shared_ptr<Page>>& pages,
- Encoding::type encoding =
Encoding::RLE_DICTIONARY) {
+inline void PaginateDict(const ColumnDescriptor* d,
+ const std::vector<typename Type::c_type>& values,
+ const std::vector<int16_t>& def_levels, int16_t
max_def_level,
+ const std::vector<int16_t>& rep_levels, int16_t
max_rep_level,
+ int num_levels_per_page, const std::vector<int>&
values_per_page,
+ std::vector<std::shared_ptr<Page>>& pages,
+ Encoding::type encoding = Encoding::RLE_DICTIONARY) {
int num_pages = static_cast<int>(values_per_page.size());
std::vector<std::shared_ptr<Buffer>> rle_indices;
std::shared_ptr<DictionaryPage> dict_page =
@@ -501,15 +497,14 @@ inline static void PaginateDict(const ColumnDescriptor* d,
// Given def/rep levels and values create multiple plain pages
template <typename Type>
-static inline void PaginatePlain(const ColumnDescriptor* d,
- const std::vector<typename Type::c_type>&
values,
- const std::vector<int16_t>& def_levels,
- int16_t max_def_level,
- const std::vector<int16_t>& rep_levels,
- int16_t max_rep_level, int
num_levels_per_page,
- const std::vector<int>& values_per_page,
- std::vector<std::shared_ptr<Page>>& pages,
- Encoding::type encoding = Encoding::PLAIN) {
+inline void PaginatePlain(const ColumnDescriptor* d,
+ const std::vector<typename Type::c_type>& values,
+ const std::vector<int16_t>& def_levels, int16_t
max_def_level,
+ const std::vector<int16_t>& rep_levels, int16_t
max_rep_level,
+ int num_levels_per_page,
+ const std::vector<int>& values_per_page,
+ std::vector<std::shared_ptr<Page>>& pages,
+ Encoding::type encoding = Encoding::PLAIN) {
int num_pages = static_cast<int>(values_per_page.size());
int def_level_start = 0;
int def_level_end = 0;
@@ -537,14 +532,12 @@ static inline void PaginatePlain(const ColumnDescriptor*
d,
// Generates pages from randomly generated data
template <typename Type>
-static inline int MakePages(const ColumnDescriptor* d, int num_pages, int
levels_per_page,
- std::vector<int16_t>& def_levels,
- std::vector<int16_t>& rep_levels,
- std::vector<typename Type::c_type>& values,
- std::vector<uint8_t>& buffer,
- std::vector<std::shared_ptr<Page>>& pages,
- Encoding::type encoding = Encoding::PLAIN,
- uint32_t seed = 0) {
+inline int MakePages(const ColumnDescriptor* d, int num_pages, int
levels_per_page,
+ std::vector<int16_t>& def_levels, std::vector<int16_t>&
rep_levels,
+ std::vector<typename Type::c_type>& values,
+ std::vector<uint8_t>& buffer,
+ std::vector<std::shared_ptr<Page>>& pages,
+ Encoding::type encoding = Encoding::PLAIN, uint32_t seed
= 0) {
int num_levels = levels_per_page * num_pages;
int num_values = 0;
int16_t zero = 0;
