[opt](build) Take olap_common.h off the dependency base and slim the PCH

be/src/common/logging.h

@@ -60,10 +60,17 @@
 
 #include <fmt/format.h>
 
-#include "util/uid_util.h"
+#include <string>
 
 namespace doris {
 
+// TaggableLogger::tag only needs these declarations; TUs that actually log a
+// TUniqueId/PUniqueId have the full types from their own includes.
+class TUniqueId;
+class PUniqueId;
+std::string print_id(const TUniqueId& id);
+std::string print_id(const PUniqueId& id);
+
 // glog doesn't allow multiple invocations of InitGoogleLogging. This method conditionally
 // calls InitGoogleLogging only if it hasn't been called before.
 bool init_glog(const char* basename);

be/src/core/column/column.h

@@ -36,7 +36,7 @@
 #include "core/string_ref.h"
 #include "core/typeid_cast.h"
 #include "core/types.h"
-#include "storage/olap_common.h"
+#include "storage/rowset_id.h"
 
 namespace doris {
 class SipHash;

be/src/core/column/column_nullable.h

@@ -31,7 +31,7 @@
 #include "core/string_ref.h"
 #include "core/typeid_cast.h"
 #include "core/types.h"
-#include "storage/olap_common.h"
+#include "storage/rowset_id.h"
 
 class SipHash;
 

be/src/core/data_type/data_type_decimal.h

@@ -46,7 +46,7 @@
 #include "core/typeid_cast.h"
 #include "core/types.h"
 #include "exec/common/arithmetic_overflow.h"
-#include "storage/olap_common.h"
+#include "storage/field_type.h"
 
 namespace doris {
 class DecimalV2Value;

be/src/core/data_type/data_type_fixed_length_object.h

@@ -31,6 +31,7 @@
 #include "core/data_type_serde/data_type_string_serde.h"
 #include "core/field.h"
 #include "core/types.h"
+#include "storage/field_type.h"
 
 namespace doris {
 

be/src/core/data_type/data_type_ipv4.h

@@ -30,7 +30,6 @@
 #include "core/data_type_serde/data_type_ipv4_serde.h"
 #include "core/pod_array.h"
 #include "core/types.h"
-#include "storage/olap_common.h"
 
 namespace doris {
 class BufferWritable;

be/src/core/data_type/data_type_nothing.h

@@ -35,6 +35,7 @@
 #include "core/data_type_serde/data_type_serde.h"
 #include "core/field.h"
 #include "core/types.h"
+#include "storage/field_type.h"
 
 namespace doris {
 

be/src/core/data_type/data_type_varbinary.h

@@ -32,6 +32,7 @@
 #include "core/data_type_serde/data_type_varbinary_serde.h"
 #include "core/field.h"
 #include "core/string_view.h"
+#include "storage/field_type.h"
 
 namespace doris {
 class BufferWritable;

be/src/core/data_type/primitive_type.h

@@ -36,6 +36,7 @@
 #include "exec/common/template_helpers.hpp"
 
 namespace doris {
+class PathInData;
 template <typename T>
 class ColumnStr;
 class IColumnDummy;

be/src/core/data_type_serde/data_type_array_serde.cpp

@@ -35,6 +35,7 @@
 #include "core/data_type_serde/orc_serde_utils.h"
 #include "core/string_ref.h"
 #include "exprs/function/function_helpers.h"
+#include "storage/field_type.h"
 #include "util/jsonb_document.h"
 #include "util/jsonb_writer.h"
 

be/src/core/data_type_serde/data_type_ipv4_serde.h

@@ -31,7 +31,6 @@
 #include "core/string_ref.h"
 #include "core/types.h"
 #include "core/value/ipv4_value.h"
-#include "storage/olap_common.h"
 
 namespace doris {
 

be/src/core/data_type_serde/data_type_ipv6_serde.h

@@ -31,7 +31,6 @@
 #include "core/string_ref.h"
 #include "core/types.h"
 #include "core/value/ipv6_value.h"
-#include "storage/olap_common.h"
 
 namespace doris {
 

be/src/core/data_type_serde/data_type_number_serde.h

@@ -29,10 +29,10 @@
 #include "core/data_type/data_type.h"
 #include "core/data_type/define_primitive_type.h"
 #include "core/data_type_serde/data_type_serde.h"
+#include "core/extended_types.h"
 #include "core/field.h"
 #include "core/string_ref.h"
 #include "core/types.h"
-#include "storage/olap_common.h"
 
 namespace doris {
 class JsonbOutStream;

be/src/core/data_type_serde/data_type_string_serde.h

@@ -28,6 +28,7 @@
 #include "core/column/column_string.h"
 #include "core/data_type_serde/data_type_serde.h"
 #include "core/types.h"
+#include "storage/field_type.h"
 
 namespace doris {
 class PValues;

be/src/core/decimal12.h

@@ -17,11 +17,15 @@
 
 #pragma once
 
+#include <cinttypes>
 #include <cstdint>
+#include <cstdio>
+#include <cstdlib>
+#include <cstring>
 #include <iostream>
 #include <string>
 
-#include "storage/utils.h"
+#include "common/status.h"
 
 namespace doris {
 
@@ -118,7 +122,7 @@ struct decimal12_t {
                     (nullptr != sepr) ? MAX_FRAC_DIGITS_NUM - static_cast<int64_t>(strlen(sepr + 1))
                                       : MAX_FRAC_DIGITS_NUM;
             frac_len = frac_len > 0 ? frac_len : 0;
-            fraction *= g_power_table[frac_len];
+            fraction *= k_power_table[frac_len];
         }
 
         if (sign != nullptr) {
@@ -132,6 +136,11 @@ struct decimal12_t {
     static const int32_t FRAC_RATIO = 1000000000;
     static const int32_t MAX_INT_DIGITS_NUM = 18;
     static const int32_t MAX_FRAC_DIGITS_NUM = 9;
+    // 10^0 .. 10^MAX_FRAC_DIGITS_NUM, used by from_string. A private copy so
+    // this header does not need storage/utils.h (whose g_power_table drags in
+    // the whole storage layer).
+    static constexpr int32_t k_power_table[MAX_FRAC_DIGITS_NUM + 1] = {
+            1, 10, 100, 1000, 10000, 100000, 1000000, 10000000, 100000000, 1000000000};
 
     int64_t integer;
     int32_t fraction;

be/src/core/extended_types.h

@@ -103,3 +103,12 @@ struct IsArithmetic {
 
 template <typename T>
 inline constexpr bool IsArithmeticV = IsArithmetic<T>::value;
+
+namespace doris {
+// Canonical 128-bit integer aliases. They live here (not in a storage header)
+// because bottom-of-the-world headers like core/packed_int128.h and
+// util/coding.h must be able to name them without dragging in the storage
+// layer; storage/olap_common.h re-exports them by including this header.
+using int128_t = __int128;
+using uint128_t = unsigned __int128;
+} // namespace doris

be/src/core/field.cpp

@@ -31,6 +31,7 @@
 #include "core/value/timestamptz_value.h"
 #include "core/value/vdatetime_value.h"
 #include "exprs/function/cast/cast_to_string.h"
+#include "util/json/path_in_data.h"
 #include "util/var_int.h"
 
 namespace doris {

be/src/core/packed_int128.h

@@ -17,7 +17,9 @@
 
 #pragma once
 
-#include "storage/olap_common.h"
+#include <cstring>
+
+#include "core/extended_types.h"
 
 namespace doris {
 

be/src/core/value/variant/variant_field.cpp

@@ -27,6 +27,7 @@
 #include "core/field.h"
 #include "core/value/variant/variant_parquet_encoding.h"
 #include "core/value/variant/variant_scalar.h"
+#include "util/json/path_in_data.h"
 #include "util/utf8_check.h"
 
 namespace doris {
@@ -285,6 +286,8 @@ void validate_variant_payload(VariantRef value) {
 VariantField::VariantField(std::unique_ptr<char[]> data, size_t size) noexcept
         : _data(std::move(data)), _size(size) {}
 
+VariantField::VariantField() noexcept = default;
+
 VariantField::~VariantField() = default;
 
 VariantField::VariantField(VariantMap legacy)

be/src/core/value/variant/variant_field.h

@@ -23,11 +23,11 @@
 
 #include "core/string_ref.h"
 #include "core/value/variant/variant_value.h"
-#include "util/json/path_in_data.h"
 
 namespace doris {
 
 struct FieldWithDataType;
+class PathInData;
 class VariantScalarRef;
 using VariantMap = std::map<PathInData, FieldWithDataType>;
 
@@ -43,7 +43,9 @@ void validate_variant_payload(VariantRef value);
 // [u32 little-endian metadata_size][metadata][exactly one value].
 class VariantField {
 public:
-    VariantField() noexcept = default;
+    // Every special member lives in the .cpp: with PathInData forward-declared, any inline
+    // definition would instantiate the VariantMap destructor through the _legacy deleter.
+    VariantField() noexcept;
     ~VariantField();
 
     VariantField(const VariantField& other);

be/src/exec/operator/schema_scan_operator.cpp

@@ -19,6 +19,7 @@
 
 #include <gen_cpp/FrontendService_types.h>
 
+#include <boost/algorithm/string.hpp>
 #include <memory>
 
 #include "core/column/column_nullable.h"

be/src/exprs/function/cast/cast_to_string.h

@@ -21,6 +21,7 @@
 #include "core/types.h"
 #include "core/value/time_value.h"
 #include "exprs/function/cast/cast_base.h"
+#include "runtime/runtime_state.h"
 #include "util/mysql_global.h"
 #include "util/to_string.h"
 namespace doris {

be/src/exprs/function/cast/cast_to_timestamptz.h

@@ -27,6 +27,7 @@
 #include "exprs/function/cast/cast_base.h"
 #include "exprs/function/cast/cast_to_datetimev2_impl.hpp"
 #include "exprs/function/cast/cast_to_timestamptz_impl.hpp"
+#include "runtime/runtime_state.h"
 
 namespace doris {
 

be/src/exprs/function/function.h

@@ -42,15 +42,22 @@
 #include "core/data_type/data_type_struct.h"
 #include "core/data_type/define_primitive_type.h"
 #include "core/types.h"
-#include "exprs/expr_zonemap_filter.h"
 #include "exprs/function_context.h"
 #include "exprs/vexpr_fwd.h"
-#include "storage/index/inverted/inverted_index_iterator.h" // IWYU pragma: keep
-#include "storage/index/inverted/inverted_index_parser.h"
 #include "storage/index/zone_map/zonemap_filter_result.h"
 
 namespace doris {
 struct InvertedIndexAnalyzerCtx;
+namespace expr_zonemap {
+struct DictionaryEvalContext;
+struct BloomFilterEvalContext;
+} // namespace expr_zonemap
+using DictionaryEvalContext = expr_zonemap::DictionaryEvalContext;
+using BloomFilterEvalContext = expr_zonemap::BloomFilterEvalContext;
+namespace segment_v2 {
+class IndexIterator;
+class InvertedIndexResultBitmap;
+} // namespace segment_v2
 } // namespace doris
 
 namespace doris {

be/src/exprs/function/function_encryption.cpp

@@ -315,18 +315,30 @@ struct EncryptionAndDecryptMultiImpl {
         auto& result_offset = result_column->get_offsets();
         result_offset.resize(input_rows_count);
 
-        if ((arg_num == 5) && col_const[1] && col_const[2] && col_const[3] && col_const[4]) {
-            vector_const(assert_cast<const ColumnString*>(argument_columns[0].get()),
-                         argument_columns[1]->get_data_at(0), argument_columns[2]->get_data_at(0),
-                         argument_columns[3]->get_data_at(0), input_rows_count, result_data,
-                         result_offset, result_null_map_column->get_data(),
-                         argument_columns[4]->get_data_at(0));
-        } else if ((arg_num == 4) && col_const[1] && col_const[2] && col_const[3]) {
-            vector_const(assert_cast<const ColumnString*>(argument_columns[0].get()),
-                         argument_columns[1]->get_data_at(0), argument_columns[2]->get_data_at(0),
-                         argument_columns[3]->get_data_at(0), input_rows_count, result_data,
-                         result_offset, result_null_map_column->get_data(), StringRef());
-        } else {
+        // if constexpr: the discarded arg_num instantiation must not index
+        // col_const[4] / argument_columns[4] out of bounds (-Warray-bounds).
+        bool all_params_const = false;
+        if constexpr (arg_num == 5) {
+            if (col_const[1] && col_const[2] && col_const[3] && col_const[4]) {
+                vector_const(assert_cast<const ColumnString*>(argument_columns[0].get()),
+                             argument_columns[1]->get_data_at(0),
+                             argument_columns[2]->get_data_at(0),
+                             argument_columns[3]->get_data_at(0), input_rows_count, result_data,
+                             result_offset, result_null_map_column->get_data(),
+                             argument_columns[4]->get_data_at(0));
+                all_params_const = true;
+            }
+        } else if constexpr (arg_num == 4) {
+            if (col_const[1] && col_const[2] && col_const[3]) {
+                vector_const(assert_cast<const ColumnString*>(argument_columns[0].get()),
+                             argument_columns[1]->get_data_at(0),
+                             argument_columns[2]->get_data_at(0),
+                             argument_columns[3]->get_data_at(0), input_rows_count, result_data,
+                             result_offset, result_null_map_column->get_data(), StringRef());
+                all_params_const = true;
+            }
+        }
+        if (!all_params_const) {
             std::vector<const ColumnString::Offsets*> offsets_list(argument_size);
             std::vector<const ColumnString::Chars*> chars_list(argument_size);
             for (size_t i = 0; i < argument_size; ++i) {

be/src/exprs/function/function_ip.h

@@ -49,6 +49,7 @@
 #include "exprs/function/function.h"
 #include "exprs/function/function_helpers.h"
 #include "storage/index/index_reader_helper.h"
+#include "storage/index/inverted/inverted_index_iterator.h"
 
 namespace doris {
 

be/src/exprs/function/functions_comparison.h

@@ -48,6 +48,7 @@
 #include "exprs/function/functions_logical.h"
 #include "exprs/vexpr.h"
 #include "storage/index/index_reader_helper.h"
+#include "storage/index/inverted/inverted_index_iterator.h"
 
 namespace doris {
 /** Comparison functions: ==, !=, <, >, <=, >=.

be/src/exprs/function/is_not_null.h

@@ -40,6 +40,7 @@
 #include "exprs/expr_zonemap_filter.h"
 #include "exprs/function/function.h"
 #include "exprs/vslot_ref.h"
+#include "storage/index/index_iterator.h"
 
 namespace doris {
 class FunctionContext;

be/src/exprs/function/is_null.h

@@ -37,6 +37,7 @@
 #include "exprs/expr_zonemap_filter.h"
 #include "exprs/function/function.h"
 #include "exprs/vslot_ref.h"
+#include "storage/index/index_iterator.h"
 
 namespace doris {
 class FunctionContext;

be/src/exprs/vmatch_predicate.cpp

@@ -29,6 +29,7 @@
 #include <gen_cpp/Exprs_types.h>
 #include <glog/logging.h>
 
+#include <boost/algorithm/string.hpp>
 #include <memory>
 #include <string>
 #include <string_view>

be/src/format_v2/file_reader.h

@@ -34,6 +34,7 @@
 #include "gen_cpp/PlanNodes_types.h"
 #include "io/file_factory.h"
 #include "io/fs/file_reader_writer_fwd.h"
+#include "io/io_common.h"
 
 namespace doris {
 class Block;

be/src/io/fs/tracing_file_reader.h

@@ -18,6 +18,7 @@
 #pragma once
 #include "common/status.h"
 #include "io/fs/file_reader.h"
+#include "io/io_common.h"
 #include "runtime/runtime_profile.h"
 
 namespace doris {

be/src/pch/pch.h

@@ -527,4 +527,3 @@
 #include "common/config.h"
 #include "common/status.h"
 #include "common/version_internal.h"
-#include "storage/olap_common.h"