Merge branch 'incbin' into no-export-dynamic

.gitmodules

@@ -340,3 +340,6 @@
 [submodule "contrib/c-ares"]
 	path = contrib/c-ares
 	url = https://github.com/c-ares/c-ares.git
+[submodule "contrib/incbin"]
+	path = contrib/incbin
+	url = https://github.com/graphitemaster/incbin.git

cmake/embed_binary.cmake

@@ -1,58 +0,0 @@
-# Embed a set of resource files into a resulting object file.
-#
-# Signature: `clickhouse_embed_binaries(TARGET <target> RESOURCE_DIR <dir> RESOURCES <resource> ...)
-#
-# This will generate a static library target named `<target>`, which contains the contents of
-# each `<resource>` file. The files should be located in `<dir>`. <dir> defaults to
-# ${CMAKE_CURRENT_SOURCE_DIR}, and the resources may not be empty.
-#
-# Each resource will result in three symbols in the final archive, based on the name `<resource>`.
-# These are:
-#   1. `_binary_<name>_start`: Points to the start of the binary data from `<resource>`.
-#   2. `_binary_<name>_end`: Points to the end of the binary data from `<resource>`.
-#   2. `_binary_<name>_size`: Points to the size of the binary data from `<resource>`.
-#
-# `<name>` is a normalized name derived from `<resource>`, by replacing the characters "./-" with
-# the character "_", and the character "+" with "_PLUS_". This scheme is similar to those generated
-# by `ld -r -b binary`, and matches the expectations in `./base/common/getResource.cpp`.
-macro(clickhouse_embed_binaries)
-    set(one_value_args TARGET RESOURCE_DIR)
-    set(resources RESOURCES)
-    cmake_parse_arguments(EMBED "" "${one_value_args}" ${resources} ${ARGN})
-
-    if (NOT DEFINED EMBED_TARGET)
-        message(FATAL_ERROR "A target name must be provided for embedding binary resources into")
-    endif()
-
-    if (NOT DEFINED EMBED_RESOURCE_DIR)
-        set(EMBED_RESOURCE_DIR "${CMAKE_CURRENT_SOURCE_DIR}")
-    endif()
-
-    list(LENGTH EMBED_RESOURCES N_RESOURCES)
-    if (N_RESOURCES LESS 1)
-        message(FATAL_ERROR "The list of binary resources to embed may not be empty")
-    endif()
-
-    add_library("${EMBED_TARGET}" STATIC)
-    set_target_properties("${EMBED_TARGET}" PROPERTIES LINKER_LANGUAGE C)
-
-    set(EMBED_TEMPLATE_FILE "${PROJECT_SOURCE_DIR}/programs/embed_binary.S.in")
-
-    foreach(RESOURCE_FILE ${EMBED_RESOURCES})
-        set(ASSEMBLY_FILE_NAME "${RESOURCE_FILE}.S")
-        set(BINARY_FILE_NAME "${RESOURCE_FILE}")
-
-        # Normalize the name of the resource.
-        string(REGEX REPLACE "[\./-]" "_" SYMBOL_NAME "${RESOURCE_FILE}") # - must be last in regex
-        string(REPLACE "+" "_PLUS_" SYMBOL_NAME "${SYMBOL_NAME}")
-
-        # Generate the configured assembly file in the output directory.
-        configure_file("${EMBED_TEMPLATE_FILE}" "${CMAKE_CURRENT_BINARY_DIR}/${ASSEMBLY_FILE_NAME}" @ONLY)
-
-        # Set the include directory for relative paths specified for `.incbin` directive.
-        set_property(SOURCE "${CMAKE_CURRENT_BINARY_DIR}/${ASSEMBLY_FILE_NAME}" APPEND PROPERTY INCLUDE_DIRECTORIES "${EMBED_RESOURCE_DIR}")
-
-        target_sources("${EMBED_TARGET}" PRIVATE "${CMAKE_CURRENT_BINARY_DIR}/${ASSEMBLY_FILE_NAME}")
-        set_target_properties("${EMBED_TARGET}" PROPERTIES OBJECT_DEPENDS "${RESOURCE_FILE}")
-    endforeach()
-endmacro()

contrib/CMakeLists.txt

@@ -164,13 +164,13 @@ add_contrib (libpq-cmake libpq)
 add_contrib (nuraft-cmake NuRaft)
 add_contrib (fast_float-cmake fast_float)
 add_contrib (datasketches-cpp-cmake datasketches-cpp)
+add_contrib (incbin-cmake incbin)
 
 option(ENABLE_NLP "Enable NLP functions support" ${ENABLE_LIBRARIES})
 if (ENABLE_NLP)
     add_contrib (libstemmer-c-cmake libstemmer_c)
     add_contrib (wordnet-blast-cmake wordnet-blast)
     add_contrib (lemmagen-c-cmake lemmagen-c)
-    add_contrib (nlp-data-cmake nlp-data)
     add_contrib (cld2-cmake cld2)
 endif()
 

contrib/cctz-cmake/CMakeLists.txt

@@ -1,4 +1,3 @@
-include(${ClickHouse_SOURCE_DIR}/cmake/embed_binary.cmake)
 set(LIBRARY_DIR "${ClickHouse_SOURCE_DIR}/contrib/cctz")
 
 set (SRCS
@@ -23,12 +22,10 @@ if (OS_FREEBSD)
 endif ()
 
 # Related to time_zones table:
-# StorageSystemTimeZones.generated.cpp is autogenerated each time during a build
+# TimeZones.generated.cpp is autogenerated each time during a build
-# data in this file will be used to populate the system.time_zones table, this is specific to OS_LINUX
+set(TIMEZONES_FILE "${CMAKE_CURRENT_BINARY_DIR}/TimeZones.generated.cpp")
-# as the library that's built using embedded tzdata is also specific to OS_LINUX
-set(SYSTEM_STORAGE_TZ_FILE "${PROJECT_BINARY_DIR}/src/Storages/System/StorageSystemTimeZones.generated.cpp")
 # remove existing copies so that its generated fresh on each build.
-file(REMOVE ${SYSTEM_STORAGE_TZ_FILE})
+file(REMOVE ${TIMEZONES_FILE})
 
 # get the list of timezones from tzdata shipped with cctz
 set(TZDIR "${LIBRARY_DIR}/testdata/zoneinfo")
@@ -36,28 +33,45 @@ file(STRINGS "${LIBRARY_DIR}/testdata/version" TZDATA_VERSION)
 set_property(GLOBAL PROPERTY TZDATA_VERSION_PROP "${TZDATA_VERSION}")
 message(STATUS "Packaging with tzdata version: ${TZDATA_VERSION}")
 
-set(TIMEZONE_RESOURCE_FILES)
-
 # each file in that dir (except of tab and localtime) store the info about timezone
 execute_process(COMMAND
     bash -c "cd ${TZDIR} && find * -type f -and ! -name '*.tab' -and ! -name 'localtime' | LC_ALL=C sort | paste -sd ';' -"
     OUTPUT_STRIP_TRAILING_WHITESPACE
     OUTPUT_VARIABLE TIMEZONES)
 
-file(APPEND ${SYSTEM_STORAGE_TZ_FILE} "// autogenerated by ClickHouse/contrib/cctz-cmake/CMakeLists.txt\n")
+file(APPEND ${TIMEZONES_FILE} "// autogenerated by ClickHouse/contrib/cctz-cmake/CMakeLists.txt\n")
-file(APPEND ${SYSTEM_STORAGE_TZ_FILE} "const char * auto_time_zones[] {\n" )
+file(APPEND ${TIMEZONES_FILE} "#include <incbin.h>\n")
+
+set (COUNTER 1)
+foreach(TIMEZONE ${TIMEZONES})
+    file(APPEND ${TIMEZONES_FILE} "INCBIN(resource_timezone${COUNTER}, \"${TIMEZONE}\");\n")
+    MATH(EXPR COUNTER "${COUNTER}+1")
+endforeach(TIMEZONE)
+
+file(APPEND ${TIMEZONES_FILE} "const char * auto_time_zones[] {\n" )
 
 foreach(TIMEZONE ${TIMEZONES})
-    file(APPEND ${SYSTEM_STORAGE_TZ_FILE} "    \"${TIMEZONE}\",\n")
+    file(APPEND ${TIMEZONES_FILE} "    \"${TIMEZONE}\",\n")
-    list(APPEND TIMEZONE_RESOURCE_FILES "${TIMEZONE}")
+    MATH(EXPR COUNTER "${COUNTER}+1")
 endforeach(TIMEZONE)
-file(APPEND ${SYSTEM_STORAGE_TZ_FILE} "    nullptr};\n")
+
-clickhouse_embed_binaries(
+file(APPEND ${TIMEZONES_FILE} "    nullptr\n};\n\n")
-    TARGET tzdata
+
-    RESOURCE_DIR "${TZDIR}"
+file(APPEND ${TIMEZONES_FILE} "#include <string_view>\n\n")
-    RESOURCES ${TIMEZONE_RESOURCE_FILES}
+file(APPEND ${TIMEZONES_FILE} "std::string_view getTimeZone(const char * name)\n{\n" )
-)
+
-add_dependencies(_cctz tzdata)
+set (COUNTER 1)
-target_link_libraries(_cctz INTERFACE "-Wl,${WHOLE_ARCHIVE} $<TARGET_FILE:tzdata> -Wl,${NO_WHOLE_ARCHIVE}")
+foreach(TIMEZONE ${TIMEZONES})
+    file(APPEND ${TIMEZONES_FILE} "    if (std::string_view(\"${TIMEZONE}\") == name) return { reinterpret_cast<const char *>(gresource_timezone${COUNTER}Data), gresource_timezone${COUNTER}Size };\n")
+    MATH(EXPR COUNTER "${COUNTER}+1")
+endforeach(TIMEZONE)
+
+file(APPEND ${TIMEZONES_FILE} "    return {};\n")
+file(APPEND ${TIMEZONES_FILE} "}\n")
+
+add_library (tzdata ${TIMEZONES_FILE})
+target_link_libraries(tzdata ch_contrib::incbin)
+target_include_directories(tzdata PRIVATE ${TZDIR})
+target_link_libraries(_cctz tzdata)
 
 add_library(ch_contrib::cctz ALIAS _cctz)

contrib/incbin

@@ -0,0 +1 @@
+Subproject commit 6e576cae5ab5810f25e2631f2e0b80cbe7dc8cbf

contrib/incbin-cmake/CMakeLists.txt

@@ -0,0 +1,4 @@
+set(LIBRARY_DIR "${ClickHouse_SOURCE_DIR}/contrib/incbin")
+add_library(_incbin INTERFACE)
+target_include_directories(_incbin SYSTEM INTERFACE ${LIBRARY_DIR})
+add_library(ch_contrib::incbin ALIAS _incbin)

contrib/nlp-data-cmake/CMakeLists.txt

@@ -1,15 +0,0 @@
-include(${ClickHouse_SOURCE_DIR}/cmake/embed_binary.cmake)
-
-set(LIBRARY_DIR "${ClickHouse_SOURCE_DIR}/contrib/nlp-data")
-
-add_library (_nlp_data INTERFACE)
-
-clickhouse_embed_binaries(
-    TARGET nlp_dictionaries
-    RESOURCE_DIR "${LIBRARY_DIR}"
-    RESOURCES charset.zst tonality_ru.zst programming.zst
-)
-
-add_dependencies(_nlp_data nlp_dictionaries)
-target_link_libraries(_nlp_data INTERFACE "-Wl,${WHOLE_ARCHIVE} $<TARGET_FILE:nlp_dictionaries> -Wl,${NO_WHOLE_ARCHIVE}")
-add_library(ch_contrib::nlp_data ALIAS _nlp_data)

docker/test/fasttest/run.sh

@@ -147,6 +147,7 @@ function clone_submodules
             contrib/simdjson
             contrib/liburing
             contrib/libfiu
+            contrib/incbin
         )
 
         git submodule sync

programs/install/CMakeLists.txt

@@ -10,3 +10,6 @@ set (CLICKHOUSE_INSTALL_LINK
 )
 
 clickhouse_program_add_library(install)
+
+# For incbin
+target_include_directories(clickhouse-install-lib PRIVATE "${CMAKE_CURRENT_SOURCE_DIR}/../server")

programs/install/Install.cpp

@@ -20,10 +20,7 @@
 #include <Common/formatReadable.h>
 #include <Common/Config/ConfigProcessor.h>
 #include <Common/OpenSSLHelpers.h>
-#include <base/hex.h>
-#include <Common/getResource.h>
 #include <base/sleep.h>
-#include <IO/ReadBufferFromFileDescriptor.h>
 #include <IO/WriteBufferFromFileDescriptor.h>
 #include <IO/ReadBufferFromFile.h>
 #include <IO/WriteBufferFromFile.h>
@@ -35,6 +32,12 @@
 
 #include <Poco/Util/XMLConfiguration.h>
 
+#include <incbin.h>
+
+/// Embedded configuration files used inside the install program
+INCBIN(resource_config_xml, "config.xml");
+INCBIN(resource_users_xml, "users.xml");
+
 
 /** This tool can be used to install ClickHouse without a deb/rpm/tgz package, having only "clickhouse" binary.
   * It also allows to avoid dependency on systemd, upstart, SysV init.
@@ -560,7 +563,7 @@ int mainEntryClickHouseInstall(int argc, char ** argv)
 
         if (!fs::exists(main_config_file))
         {
-            std::string_view main_config_content = getResource("config.xml");
+            std::string_view main_config_content(reinterpret_cast<const char *>(gresource_config_xmlData), gresource_config_xmlSize);
             if (main_config_content.empty())
             {
                 fmt::print("There is no default config.xml, you have to download it and place to {}.\n", main_config_file.string());
@@ -672,7 +675,7 @@ int mainEntryClickHouseInstall(int argc, char ** argv)
 
         if (!fs::exists(users_config_file))
         {
-            std::string_view users_config_content = getResource("users.xml");
+            std::string_view users_config_content(reinterpret_cast<const char *>(gresource_users_xmlData), gresource_users_xmlSize);
             if (users_config_content.empty())
             {
                 fmt::print("There is no default users.xml, you have to download it and place to {}.\n", users_config_file.string());

programs/keeper/CMakeLists.txt

@@ -1,16 +1,3 @@
-include(${ClickHouse_SOURCE_DIR}/cmake/embed_binary.cmake)
-
-if (OS_LINUX)
-    set (LINK_RESOURCE_LIB INTERFACE "-Wl,${WHOLE_ARCHIVE} $<TARGET_FILE:clickhouse_keeper_configs> -Wl,${NO_WHOLE_ARCHIVE}")
-    # for some reason INTERFACE linkage doesn't work for standalone binary
-    set (LINK_RESOURCE_LIB_STANDALONE_KEEPER "-Wl,${WHOLE_ARCHIVE} $<TARGET_FILE:clickhouse_keeper_configs> -Wl,${NO_WHOLE_ARCHIVE}")
-endif ()
-
-clickhouse_embed_binaries(
-    TARGET clickhouse_keeper_configs
-    RESOURCES keeper_config.xml keeper_embedded.xml
-)
-
 set(CLICKHOUSE_KEEPER_SOURCES
     Keeper.cpp
 )
@@ -29,7 +16,6 @@ set (CLICKHOUSE_KEEPER_LINK
 clickhouse_program_add(keeper)
 
 install(FILES keeper_config.xml DESTINATION "${CLICKHOUSE_ETC_DIR}/clickhouse-keeper" COMPONENT clickhouse-keeper)
-add_dependencies(clickhouse-keeper-lib clickhouse_keeper_configs)
 
 if (BUILD_STANDALONE_KEEPER)
     # Straight list of all required sources
@@ -215,7 +201,6 @@ if (BUILD_STANDALONE_KEEPER)
         ${LINK_RESOURCE_LIB_STANDALONE_KEEPER}
     )
 
-    add_dependencies(clickhouse-keeper clickhouse_keeper_configs)
     set_target_properties(clickhouse-keeper PROPERTIES RUNTIME_OUTPUT_DIRECTORY ../)
 
     if (SPLIT_DEBUG_SYMBOLS)

programs/keeper/Keeper.cpp

@@ -457,8 +457,10 @@ try
     const std::string key_path = config().getString("openSSL.server.privateKeyFile", "");
 
     std::vector<std::string> extra_paths = {include_from_path};
-    if (!cert_path.empty()) extra_paths.emplace_back(cert_path);
+    if (!cert_path.empty())
-    if (!key_path.empty()) extra_paths.emplace_back(key_path);
+        extra_paths.emplace_back(cert_path);
+    if (!key_path.empty())
+        extra_paths.emplace_back(key_path);
 
     /// ConfigReloader have to strict parameters which are redundant in our case
     auto main_config_reloader = std::make_unique<ConfigReloader>(

programs/server/CMakeLists.txt

@@ -1,12 +1,8 @@
-include(${ClickHouse_SOURCE_DIR}/cmake/embed_binary.cmake)
-
 set(CLICKHOUSE_SERVER_SOURCES
     MetricsTransmitter.cpp
     Server.cpp
 )
 
-set (LINK_RESOURCE_LIB INTERFACE "-Wl,${WHOLE_ARCHIVE} $<TARGET_FILE:clickhouse_server_configs> -Wl,${NO_WHOLE_ARCHIVE}")
-
 set (CLICKHOUSE_SERVER_LINK
     PRIVATE
         clickhouse_aggregate_functions
@@ -33,10 +29,4 @@ endif()
 
 clickhouse_program_add(server)
 
-install(FILES config.xml users.xml DESTINATION "${CLICKHOUSE_ETC_DIR}/clickhouse-server" COMPONENT clickhouse)
+target_include_directories(clickhouse-server-lib PRIVATE ${CMAKE_CURRENT_SOURCE_DIR})
-
-clickhouse_embed_binaries(
-    TARGET clickhouse_server_configs
-    RESOURCES config.xml users.xml embedded.xml play.html dashboard.html js/uplot.js
-)
-add_dependencies(clickhouse-server-lib clickhouse_server_configs)

programs/server/Server.cpp

@@ -128,6 +128,10 @@
 #   include <azure/storage/common/internal/xml_wrapper.hpp>
 #endif
 
+#include <incbin.h>
+/// A minimal file used when the server is run without installation
+INCBIN(resource_embedded_xml, "embedded.xml");
+
 namespace CurrentMetrics
 {
     extern const Metric Revision;
@@ -393,6 +397,7 @@ int Server::run()
 
 void Server::initialize(Poco::Util::Application & self)
 {
+    ConfigProcessor::registerEmbeddedConfig("config.xml", std::string_view(reinterpret_cast<const char *>(gresource_embedded_xmlData), gresource_embedded_xmlSize));
     BaseDaemon::initialize(self);
     logger().information("starting up");
 
@@ -1105,8 +1110,10 @@ try
     const std::string key_path = config().getString("openSSL.server.privateKeyFile", "");
 
     std::vector<std::string> extra_paths = {include_from_path};
-    if (!cert_path.empty()) extra_paths.emplace_back(cert_path);
+    if (!cert_path.empty())
-    if (!key_path.empty()) extra_paths.emplace_back(key_path);
+        extra_paths.emplace_back(cert_path);
+    if (!key_path.empty())
+        extra_paths.emplace_back(key_path);
 
     auto main_config_reloader = std::make_unique<ConfigReloader>(
         config_path,

src/CMakeLists.txt

@@ -210,7 +210,7 @@ if (TARGET ch_contrib::jemalloc)
     target_link_libraries (clickhouse_storages_system PRIVATE ch_contrib::jemalloc)
 endif()
 
-target_link_libraries (clickhouse_common_io PUBLIC ch_contrib::sparsehash)
+target_link_libraries (clickhouse_common_io PUBLIC ch_contrib::sparsehash ch_contrib::incbin)
 
 add_subdirectory(Access/Common)
 add_subdirectory(Common/ZooKeeper)
@@ -296,7 +296,7 @@ macro (dbms_target_include_directories)
     endforeach ()
 endmacro ()
 
-dbms_target_include_directories (PUBLIC "${ClickHouse_SOURCE_DIR}/src" "${ClickHouse_BINARY_DIR}/src")
+dbms_target_include_directories (PUBLIC "${ClickHouse_SOURCE_DIR}/src" "${ClickHouse_BINARY_DIR}/src" "${ClickHouse_SOURCE_DIR}/programs/server")
 target_include_directories (clickhouse_common_io PUBLIC "${ClickHouse_SOURCE_DIR}/src" "${ClickHouse_BINARY_DIR}/src")
 
 if (TARGET ch_contrib::llvm)
@@ -561,7 +561,7 @@ if (ENABLE_NLP)
     dbms_target_link_libraries (PUBLIC ch_contrib::stemmer)
     dbms_target_link_libraries (PUBLIC ch_contrib::wnb)
     dbms_target_link_libraries (PUBLIC ch_contrib::lemmagen)
-    dbms_target_link_libraries (PUBLIC ch_contrib::nlp_data)
+    target_include_directories(clickhouse_common_io PUBLIC ${CMAKE_SOURCE_DIR}/contrib/nlp-data)
 endif()
 
 if (TARGET ch_contrib::ulid)

src/Common/CMakeLists.txt

@@ -9,5 +9,5 @@ if (ENABLE_EXAMPLES)
 endif()
 
 if (ENABLE_MYSQL)
-    add_subdirectory (mysqlxx)
+    add_subdirectory(mysqlxx)
 endif ()

src/Common/Config/ConfigProcessor.cpp

@@ -19,7 +19,6 @@
 #include <Common/ZooKeeper/KeeperException.h>
 #include <Common/StringUtils/StringUtils.h>
 #include <Common/Exception.h>
-#include <Common/getResource.h>
 #include <Common/XMLUtils.h>
 #include <Common/logger_useful.h>
 #include <base/errnoToString.h>
@@ -83,6 +82,13 @@ ConfigProcessor::~ConfigProcessor()
         Poco::Logger::destroy("ConfigProcessor");
 }
 
+static std::unordered_map<std::string, std::string_view> embedded_configs;
+
+void ConfigProcessor::registerEmbeddedConfig(std::string name, std::string_view content)
+{
+    embedded_configs[name] = content;
+}
+
 
 /// Vector containing the name of the element and a sorted list of attribute names and values
 /// (except "remove" and "replace" attributes).
@@ -281,15 +287,15 @@ void ConfigProcessor::doIncludesRecursive(
         {
             std::string value = node->nodeValue();
 
-            bool replace_occured = false;
+            bool replace_occurred = false;
             size_t pos;
             while ((pos = value.find(substitution.first)) != std::string::npos)
             {
                 value.replace(pos, substitution.first.length(), substitution.second);
-                replace_occured = true;
+                replace_occurred = true;
             }
 
-            if (replace_occured)
+            if (replace_occurred)
                 node->setNodeValue(value);
         }
     }
@@ -528,26 +534,14 @@ XMLDocumentPtr ConfigProcessor::processConfig(
     }
     else
     {
-        /// These embedded files added during build with some cmake magic.
+        /// When we can use a config embedded in the binary.
-        /// Look at the end of programs/server/CMakeLists.txt.
+        if (auto it = embedded_configs.find(path); it != embedded_configs.end())
-        std::string embedded_name;
-        if (path == "config.xml")
-            embedded_name = "embedded.xml";
-
-        if (path == "keeper_config.xml")
-            embedded_name = "keeper_embedded.xml";
-
-        /// When we can use config embedded in binary.
-        if (!embedded_name.empty())
         {
-            auto resource = getResource(embedded_name);
-            if (resource.empty())
-                throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "Configuration file {} doesn't exist and there is no embedded config", path);
             LOG_DEBUG(log, "There is no file '{}', will use embedded config.", path);
-            config = dom_parser.parseMemory(resource.data(), resource.size());
+            config = dom_parser.parseMemory(it->second.data(), it->second.size());
         }
         else
-            throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "Configuration file {} doesn't exist", path);
+            throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "Configuration file {} doesn't exist and there is no embedded config", path);
     }
 
     std::vector<std::string> contributing_files;

src/Common/Config/ConfigProcessor.h

@@ -65,6 +65,9 @@ public:
         zkutil::ZooKeeperNodeCache * zk_node_cache = nullptr,
         const zkutil::EventPtr & zk_changed_event = nullptr);
 
+    /// These configurations will be used if there is no configuration file.
+    static void registerEmbeddedConfig(std::string name, std::string_view content);
+
 
     /// loadConfig* functions apply processConfig and create Poco::Util::XMLConfiguration.
     /// The resulting XML document is saved into a file with the name

src/Common/DateLUTImpl.cpp

@@ -3,7 +3,6 @@
 #include <cctz/civil_time.h>
 #include <cctz/time_zone.h>
 #include <cctz/zone_info_source.h>
-#include <Common/getResource.h>
 #include <Poco/Exception.h>
 
 #include <algorithm>
@@ -11,6 +10,11 @@
 #include <chrono>
 #include <cstring>
 #include <memory>
+#include <iostream>
+
+
+/// Embedded timezones.
+std::string_view getTimeZone(const char * name);
 
 
 namespace
@@ -249,9 +253,10 @@ namespace cctz_extension
             const std::string & name,
             const std::function<std::unique_ptr<cctz::ZoneInfoSource>(const std::string & name)> & fallback)
         {
-            std::string_view resource = getResource(name);
+            std::string_view tz_file = getTimeZone(name.data());
-            if (!resource.empty())
+
-                return std::make_unique<Source>(resource.data(), resource.size());
+            if (!tz_file.empty())
+                return std::make_unique<Source>(tz_file.data(), tz_file.size());
 
             return fallback(name);
         }

src/Common/FrequencyHolder.cpp

@@ -0,0 +1,185 @@
+#include <Common/FrequencyHolder.h>
+
+#if USE_NLP
+
+#include <incbin.h>
+
+/// Embedded SQL definitions
+INCBIN(resource_charset_zst, "charset.zst");
+INCBIN(resource_tonality_ru_zst, "tonality_ru.zst");
+INCBIN(resource_programming_zst, "programming.zst");
+
+
+namespace DB
+{
+
+namespace ErrorCodes
+{
+    extern const int FILE_DOESNT_EXIST;
+}
+
+
+FrequencyHolder & FrequencyHolder::getInstance()
+{
+    static FrequencyHolder instance;
+    return instance;
+}
+
+FrequencyHolder::FrequencyHolder()
+{
+    loadEmotionalDict();
+    loadEncodingsFrequency();
+    loadProgrammingFrequency();
+}
+
+void FrequencyHolder::loadEncodingsFrequency()
+{
+    Poco::Logger * log = &Poco::Logger::get("EncodingsFrequency");
+
+    LOG_TRACE(log, "Loading embedded charset frequencies");
+
+    std::string_view resource(reinterpret_cast<const char *>(gresource_charset_zstData), gresource_charset_zstSize);
+    if (resource.empty())
+        throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "There is no embedded charset frequencies");
+
+    String line;
+    UInt16 bigram;
+    Float64 frequency;
+    String charset_name;
+
+    auto buf = std::make_unique<ReadBufferFromMemory>(resource.data(), resource.size());
+    ZstdInflatingReadBuffer in(std::move(buf));
+
+    while (!in.eof())
+    {
+        readString(line, in);
+        in.ignore();
+
+        if (line.empty())
+            continue;
+
+        ReadBufferFromString buf_line(line);
+
+        // Start loading a new charset
+        if (line.starts_with("// "))
+        {
+            // Skip "// "
+            buf_line.ignore(3);
+            readString(charset_name, buf_line);
+
+            /* In our dictionary we have lines with form: <Language>_<Charset>
+            * If we need to find language of data, we return <Language>
+            * If we need to find charset of data, we return <Charset>.
+            */
+            size_t sep = charset_name.find('_');
+
+            Encoding enc;
+            enc.lang = charset_name.substr(0, sep);
+            enc.name = charset_name.substr(sep + 1);
+            encodings_freq.push_back(std::move(enc));
+        }
+        else
+        {
+            readIntText(bigram, buf_line);
+            buf_line.ignore();
+            readFloatText(frequency, buf_line);
+
+            encodings_freq.back().map[bigram] = frequency;
+        }
+    }
+    LOG_TRACE(log, "Charset frequencies was added, charsets count: {}", encodings_freq.size());
+}
+
+void FrequencyHolder::loadEmotionalDict()
+{
+    Poco::Logger * log = &Poco::Logger::get("EmotionalDict");
+    LOG_TRACE(log, "Loading embedded emotional dictionary");
+
+    std::string_view resource(reinterpret_cast<const char *>(gresource_tonality_ru_zstData), gresource_tonality_ru_zstSize);
+    if (resource.empty())
+        throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "There is no embedded emotional dictionary");
+
+    String line;
+    String word;
+    Float64 tonality;
+    size_t count = 0;
+
+    auto buf = std::make_unique<ReadBufferFromMemory>(resource.data(), resource.size());
+    ZstdInflatingReadBuffer in(std::move(buf));
+
+    while (!in.eof())
+    {
+        readString(line, in);
+        in.ignore();
+
+        if (line.empty())
+            continue;
+
+        ReadBufferFromString buf_line(line);
+
+        readStringUntilWhitespace(word, buf_line);
+        buf_line.ignore();
+        readFloatText(tonality, buf_line);
+
+        StringRef ref{string_pool.insert(word.data(), word.size()), word.size()};
+        emotional_dict[ref] = tonality;
+        ++count;
+    }
+    LOG_TRACE(log, "Emotional dictionary was added. Word count: {}", std::to_string(count));
+}
+
+void FrequencyHolder::loadProgrammingFrequency()
+{
+    Poco::Logger * log = &Poco::Logger::get("ProgrammingFrequency");
+
+    LOG_TRACE(log, "Loading embedded programming languages frequencies loading");
+
+    std::string_view resource(reinterpret_cast<const char *>(gresource_programming_zstData), gresource_programming_zstSize);
+    if (resource.empty())
+        throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "There is no embedded programming languages frequencies");
+
+    String line;
+    String bigram;
+    Float64 frequency;
+    String programming_language;
+
+    auto buf = std::make_unique<ReadBufferFromMemory>(resource.data(), resource.size());
+    ZstdInflatingReadBuffer in(std::move(buf));
+
+    while (!in.eof())
+    {
+        readString(line, in);
+        in.ignore();
+
+        if (line.empty())
+            continue;
+
+        ReadBufferFromString buf_line(line);
+
+        // Start loading a new language
+        if (line.starts_with("// "))
+        {
+            // Skip "// "
+            buf_line.ignore(3);
+            readString(programming_language, buf_line);
+
+            Language lang;
+            lang.name = programming_language;
+            programming_freq.push_back(std::move(lang));
+        }
+        else
+        {
+            readStringUntilWhitespace(bigram, buf_line);
+            buf_line.ignore();
+            readFloatText(frequency, buf_line);
+
+            StringRef ref{string_pool.insert(bigram.data(), bigram.size()), bigram.size()};
+            programming_freq.back().map[ref] = frequency;
+        }
+    }
+    LOG_TRACE(log, "Programming languages frequencies was added");
+}
+
+}
+
+#endif

src/Common/FrequencyHolder.h

@@ -1,5 +1,9 @@
 #pragma once
 
+#include "config.h"
+
+#if USE_NLP
+
 #include <base/StringRef.h>
 #include <Common/logger_useful.h>
 
@@ -7,7 +11,6 @@
 #include <unordered_map>
 
 #include <Common/Arena.h>
-#include <Common/getResource.h>
 #include <Common/HashTable/HashMap.h>
 #include <Common/StringUtils/StringUtils.h>
 #include <IO/ReadBufferFromFile.h>
@@ -20,11 +23,6 @@
 namespace DB
 {
 
-namespace ErrorCodes
-{
-    extern const int FILE_DOESNT_EXIST;
-}
-
 /// FrequencyHolder class is responsible for storing and loading dictionaries
 /// needed for text classification functions:
 ///
@@ -56,11 +54,7 @@ public:
     using EncodingMap = HashMap<UInt16, Float64>;
     using EncodingContainer = std::vector<Encoding>;
 
-    static FrequencyHolder & getInstance()
+    static FrequencyHolder & getInstance();
-    {
-        static FrequencyHolder instance;
-        return instance;
-    }
 
     const Map & getEmotionalDict() const
     {
@@ -78,161 +72,11 @@ public:
     }
 
 private:
+    FrequencyHolder();
 
-    FrequencyHolder()
+    void loadEncodingsFrequency();
-    {
+    void loadEmotionalDict();
-        loadEmotionalDict();
+    void loadProgrammingFrequency();
-        loadEncodingsFrequency();
-        loadProgrammingFrequency();
-    }
-
-    void loadEncodingsFrequency()
-    {
-        Poco::Logger * log = &Poco::Logger::get("EncodingsFrequency");
-
-        LOG_TRACE(log, "Loading embedded charset frequencies");
-
-        auto resource = getResource("charset.zst");
-            if (resource.empty())
-                throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "There is no embedded charset frequencies");
-
-        String line;
-        UInt16 bigram;
-        Float64 frequency;
-        String charset_name;
-
-        auto buf = std::make_unique<ReadBufferFromMemory>(resource.data(), resource.size());
-        ZstdInflatingReadBuffer in(std::move(buf));
-
-        while (!in.eof())
-        {
-            readString(line, in);
-            in.ignore();
-
-            if (line.empty())
-                continue;
-
-            ReadBufferFromString buf_line(line);
-
-            // Start loading a new charset
-            if (line.starts_with("// "))
-            {
-                // Skip "// "
-                buf_line.ignore(3);
-                readString(charset_name, buf_line);
-
-                /* In our dictionary we have lines with form: <Language>_<Charset>
-                * If we need to find language of data, we return <Language>
-                * If we need to find charset of data, we return <Charset>.
-                */
-                size_t sep = charset_name.find('_');
-
-                Encoding enc;
-                enc.lang = charset_name.substr(0, sep);
-                enc.name = charset_name.substr(sep + 1);
-                encodings_freq.push_back(std::move(enc));
-            }
-            else
-            {
-                readIntText(bigram, buf_line);
-                buf_line.ignore();
-                readFloatText(frequency, buf_line);
-
-                encodings_freq.back().map[bigram] = frequency;
-            }
-        }
-        LOG_TRACE(log, "Charset frequencies was added, charsets count: {}", encodings_freq.size());
-    }
-
-    void loadEmotionalDict()
-    {
-        Poco::Logger * log = &Poco::Logger::get("EmotionalDict");
-        LOG_TRACE(log, "Loading embedded emotional dictionary");
-
-        auto resource = getResource("tonality_ru.zst");
-            if (resource.empty())
-                throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "There is no embedded emotional dictionary");
-
-        String line;
-        String word;
-        Float64 tonality;
-        size_t count = 0;
-
-        auto buf = std::make_unique<ReadBufferFromMemory>(resource.data(), resource.size());
-        ZstdInflatingReadBuffer in(std::move(buf));
-
-        while (!in.eof())
-        {
-            readString(line, in);
-            in.ignore();
-
-            if (line.empty())
-                continue;
-
-            ReadBufferFromString buf_line(line);
-
-            readStringUntilWhitespace(word, buf_line);
-            buf_line.ignore();
-            readFloatText(tonality, buf_line);
-
-            StringRef ref{string_pool.insert(word.data(), word.size()), word.size()};
-            emotional_dict[ref] = tonality;
-            ++count;
-        }
-        LOG_TRACE(log, "Emotional dictionary was added. Word count: {}", std::to_string(count));
-    }
-
-    void loadProgrammingFrequency()
-    {
-        Poco::Logger * log = &Poco::Logger::get("ProgrammingFrequency");
-
-        LOG_TRACE(log, "Loading embedded programming languages frequencies loading");
-
-        auto resource = getResource("programming.zst");
-            if (resource.empty())
-                throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "There is no embedded programming languages frequencies");
-
-        String line;
-        String bigram;
-        Float64 frequency;
-        String programming_language;
-
-        auto buf = std::make_unique<ReadBufferFromMemory>(resource.data(), resource.size());
-        ZstdInflatingReadBuffer in(std::move(buf));
-
-        while (!in.eof())
-        {
-            readString(line, in);
-            in.ignore();
-
-            if (line.empty())
-                continue;
-
-            ReadBufferFromString buf_line(line);
-
-            // Start loading a new language
-            if (line.starts_with("// "))
-            {
-                // Skip "// "
-                buf_line.ignore(3);
-                readString(programming_language, buf_line);
-
-                Language lang;
-                lang.name = programming_language;
-                programming_freq.push_back(std::move(lang));
-            }
-            else
-            {
-                readStringUntilWhitespace(bigram, buf_line);
-                buf_line.ignore();
-                readFloatText(frequency, buf_line);
-
-                StringRef ref{string_pool.insert(bigram.data(), bigram.size()), bigram.size()};
-                programming_freq.back().map[ref] = frequency;
-            }
-        }
-        LOG_TRACE(log, "Programming languages frequencies was added");
-    }
 
     Arena string_pool;
 
@@ -241,3 +85,5 @@ private:
     EncodingContainer encodings_freq;
 };
 }
+
+#endif

src/Common/SymbolIndex.cpp

@@ -88,50 +88,13 @@ namespace
 /// https://stackoverflow.com/questions/32088140/multiple-string-tables-in-elf-object
 
 
-void updateResources(ElfW(Addr) base_address, std::string_view object_name, std::string_view name, const void * address, SymbolIndex::Resources & resources)
-{
-    const char * char_address = static_cast<const char *>(address);
-
-    if (name.starts_with("_binary_") || name.starts_with("binary_"))
-    {
-        if (name.ends_with("_start"))
-        {
-            name = name.substr((name[0] == '_') + strlen("binary_"));
-            name = name.substr(0, name.size() - strlen("_start"));
-
-            auto & resource = resources[name];
-            if (!resource.base_address || resource.base_address == base_address)
-            {
-                resource.base_address = base_address;
-                resource.start = std::string_view{char_address, 0}; // NOLINT(bugprone-string-constructor)
-                resource.object_name = object_name;
-            }
-        }
-        if (name.ends_with("_end"))
-        {
-            name = name.substr((name[0] == '_') + strlen("binary_"));
-            name = name.substr(0, name.size() - strlen("_end"));
-
-            auto & resource = resources[name];
-            if (!resource.base_address || resource.base_address == base_address)
-            {
-                resource.base_address = base_address;
-                resource.end = std::string_view{char_address, 0}; // NOLINT(bugprone-string-constructor)
-                resource.object_name = object_name;
-            }
-        }
-    }
-}
-
-
 /// Based on the code of musl-libc and the answer of Kanalpiroge on
 /// https://stackoverflow.com/questions/15779185/list-all-the-functions-symbols-on-the-fly-in-c-code-on-a-linux-architecture
 /// It does not extract all the symbols (but only public - exported and used for dynamic linking),
 /// but will work if we cannot find or parse ELF files.
 void collectSymbolsFromProgramHeaders(
     dl_phdr_info * info,
-    std::vector<SymbolIndex::Symbol> & symbols,
+    std::vector<SymbolIndex::Symbol> & symbols)
-    SymbolIndex::Resources & resources)
 {
     /* Iterate over all headers of the current shared lib
      * (first call is for the executable itself)
@@ -265,9 +228,6 @@ void collectSymbolsFromProgramHeaders(
                     /// We are not interested in empty symbols.
                     if (elf_sym[sym_index].st_size)
                         symbols.push_back(symbol);
-
-                    /// But resources can be represented by a pair of empty symbols (indicating their boundaries).
-                    updateResources(base_address, info->dlpi_name, symbol.name, symbol.address_begin, resources);
                 }
 
                 break;
@@ -303,8 +263,7 @@ void collectSymbolsFromELFSymbolTable(
     const Elf & elf,
     const Elf::Section & symbol_table,
     const Elf::Section & string_table,
-    std::vector<SymbolIndex::Symbol> & symbols,
+    std::vector<SymbolIndex::Symbol> & symbols)
-    SymbolIndex::Resources & resources)
 {
     /// Iterate symbol table.
     const ElfSym * symbol_table_entry = reinterpret_cast<const ElfSym *>(symbol_table.begin());
@@ -334,8 +293,6 @@ void collectSymbolsFromELFSymbolTable(
 
         if (symbol_table_entry->st_size)
             symbols.push_back(symbol);
-
-        updateResources(info->dlpi_addr, info->dlpi_name, symbol.name, symbol.address_begin, resources);
     }
 }
 
@@ -345,8 +302,7 @@ bool searchAndCollectSymbolsFromELFSymbolTable(
     const Elf & elf,
     unsigned section_header_type,
     const char * string_table_name,
-    std::vector<SymbolIndex::Symbol> & symbols,
+    std::vector<SymbolIndex::Symbol> & symbols)
-    SymbolIndex::Resources & resources)
 {
     std::optional<Elf::Section> symbol_table;
     std::optional<Elf::Section> string_table;
@@ -364,7 +320,7 @@ bool searchAndCollectSymbolsFromELFSymbolTable(
         return false;
     }
 
-    collectSymbolsFromELFSymbolTable(info, elf, *symbol_table, *string_table, symbols, resources);
+    collectSymbolsFromELFSymbolTable(info, elf, *symbol_table, *string_table, symbols);
     return true;
 }
 
@@ -373,7 +329,6 @@ void collectSymbolsFromELF(
     dl_phdr_info * info,
     std::vector<SymbolIndex::Symbol> & symbols,
     std::vector<SymbolIndex::Object> & objects,
-    SymbolIndex::Resources & resources,
     String & build_id)
 {
     String object_name;
@@ -485,11 +440,11 @@ void collectSymbolsFromELF(
     object.name = object_name;
     objects.push_back(std::move(object));
 
-    searchAndCollectSymbolsFromELFSymbolTable(info, *objects.back().elf, SHT_SYMTAB, ".strtab", symbols, resources);
+    searchAndCollectSymbolsFromELFSymbolTable(info, *objects.back().elf, SHT_SYMTAB, ".strtab", symbols);
 
     /// Unneeded if they were parsed from "program headers" of loaded objects.
 #if defined USE_MUSL
-    searchAndCollectSymbolsFromELFSymbolTable(info, *objects.back().elf, SHT_DYNSYM, ".dynstr", symbols, resources);
+    searchAndCollectSymbolsFromELFSymbolTable(info, *objects.back().elf, SHT_DYNSYM, ".dynstr", symbols);
 #endif
 }
 
@@ -502,8 +457,8 @@ int collectSymbols(dl_phdr_info * info, size_t, void * data_ptr)
 {
     SymbolIndex::Data & data = *reinterpret_cast<SymbolIndex::Data *>(data_ptr);
 
-    collectSymbolsFromProgramHeaders(info, data.symbols, data.resources);
+    collectSymbolsFromProgramHeaders(info, data.symbols);
-    collectSymbolsFromELF(info, data.symbols, data.objects, data.resources, data.build_id);
+    collectSymbolsFromELF(info, data.symbols, data.objects, data.build_id);
 
     /* Continue iterations */
     return 0;

src/Common/SymbolIndex.h

@@ -8,6 +8,7 @@
 #include <Common/Elf.h>
 #include <boost/noncopyable.hpp>
 
+
 namespace DB
 {
 
@@ -45,44 +46,15 @@ public:
     const std::vector<Symbol> & symbols() const { return data.symbols; }
     const std::vector<Object> & objects() const { return data.objects; }
 
-    std::string_view getResource(String name) const
-    {
-        if (auto it = data.resources.find(name); it != data.resources.end())
-            return it->second.data();
-        return {};
-    }
-
     /// The BuildID that is generated by compiler.
     String getBuildID() const { return data.build_id; }
     String getBuildIDHex() const;
 
-    struct ResourcesBlob
-    {
-        /// Symbol can be presented in multiple shared objects,
-        /// base_address will be used to compare only symbols from the same SO.
-        ElfW(Addr) base_address = 0;
-        /// Just a human name of the SO.
-        std::string_view object_name;
-        /// Data blob.
-        std::string_view start;
-        std::string_view end;
-
-        std::string_view data() const
-        {
-            assert(end.data() >= start.data());
-            return std::string_view{start.data(), static_cast<size_t>(end.data() - start.data())};
-        }
-    };
-    using Resources = std::unordered_map<std::string_view /* symbol name */, ResourcesBlob>;
-
     struct Data
     {
         std::vector<Symbol> symbols;
         std::vector<Object> objects;
         String build_id;
-
-        /// Resources (embedded binary data) are located by symbols in form of _binary_name_start and _binary_name_end.
-        Resources resources;
     };
 private:
     Data data;

src/Common/getResource.cpp

@@ -1,17 +0,0 @@
-#include "getResource.h"
-#include <boost/algorithm/string/replace.hpp>
-#include <Common/SymbolIndex.h>
-
-
-std::string_view getResource(std::string_view name)
-{
-    // Convert the resource file name into the form generated by `ld -r -b binary`.
-    std::string name_replaced(name);
-    std::replace(name_replaced.begin(), name_replaced.end(), '/', '_');
-    std::replace(name_replaced.begin(), name_replaced.end(), '-', '_');
-    std::replace(name_replaced.begin(), name_replaced.end(), '.', '_');
-    boost::replace_all(name_replaced, "+", "_PLUS_");
-
-    /// If static linking is used, we cannot use dlsym and have to parse ELF symbol table by ourself.
-    return DB::SymbolIndex::instance().getResource(name_replaced);
-}

src/Common/getResource.h

@@ -1,7 +0,0 @@
-#pragma once
-
-#include <string_view>
-
-/// Get resource from binary if exists. Otherwise return empty string view.
-/// Resources are data that is embedded into executable at link time.
-std::string_view getResource(std::string_view name);

src/Common/tests/gtest_DateLUTImpl.cpp

@@ -548,4 +548,3 @@ INSTANTIATE_TEST_SUITE_P(AllTimezones_Year1970,
 //            {0, 0 + 11 * 3600 * 24 + 12, 11},
         }))
 );
-

src/Daemon/BaseDaemon.cpp

@@ -38,7 +38,6 @@
 #include <base/coverage.h>
 #include <base/sleep.h>
 
-#include <IO/WriteBufferFromFile.h>
 #include <IO/WriteBufferFromFileDescriptorDiscardOnFailure.h>
 #include <IO/ReadBufferFromFileDescriptor.h>
 #include <IO/ReadHelpers.h>

src/Functions/FunctionsCharsetClassification.cpp

@@ -1,9 +1,12 @@
 #include <Common/FrequencyHolder.h>
+
+#if USE_NLP
+
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsTextClassification.h>
 
 #include <memory>
-#include <unordered_map>
+
 
 namespace DB
 {
@@ -46,7 +49,7 @@ namespace
         return res;
     }
 
-    /// Сount how many times each bigram occurs in the text.
+    /// Count how many times each bigram occurs in the text.
     template <typename ModelMap>
     ALWAYS_INLINE inline void calculateStats(
         const UInt8 * data,
@@ -150,3 +153,5 @@ REGISTER_FUNCTION(DetectCharset)
 }
 
 }
+
+#endif

src/Functions/FunctionsLanguageClassification.cpp

@@ -5,19 +5,17 @@
 #include <Columns/ColumnMap.h>
 #include <Columns/ColumnArray.h>
 #include <Columns/ColumnString.h>
-#include <Columns/ColumnsNumber.h>
 #include <Common/isValidUTF8.h>
 #include <DataTypes/DataTypeMap.h>
 #include <DataTypes/DataTypeString.h>
-#include <DataTypes/DataTypeTuple.h>
 #include <DataTypes/DataTypesNumber.h>
 #include <Functions/FunctionHelpers.h>
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsTextClassification.h>
-#include <Interpreters/Context.h>
 
 #include <compact_lang_det.h>
 
+
 namespace DB
 {
 /* Determine language of Unicode UTF-8 text.

src/Functions/FunctionsProgrammingClassification.cpp

@@ -1,4 +1,7 @@
 #include <Common/FrequencyHolder.h>
+
+#if USE_NLP
+
 #include <Common/StringUtils/StringUtils.h>
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsTextClassification.h>
@@ -118,3 +121,5 @@ REGISTER_FUNCTION(DetectProgrammingLanguage)
 }
 
 }
+
+#endif

src/Functions/FunctionsTonalityClassification.cpp

@@ -1,4 +1,7 @@
 #include <Common/FrequencyHolder.h>
+
+#if USE_NLP
+
 #include <Common/StringUtils/StringUtils.h>
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsTextClassification.h>
@@ -87,3 +90,5 @@ REGISTER_FUNCTION(DetectTonality)
 }
 
 }
+
+#endif

src/Server/WebUIRequestHandler.cpp

@@ -6,10 +6,16 @@
 #include <Poco/Util/LayeredConfiguration.h>
 
 #include <IO/HTTPCommon.h>
-#include <Common/getResource.h>
 
 #include <re2/re2.h>
 
+#include <incbin.h>
+
+/// Embedded HTML pages
+INCBIN(resource_play_html, "play.html");
+INCBIN(resource_dashboard_html, "dashboard.html");
+INCBIN(resource_uplot_js, "js/uplot.js");
+
 
 namespace DB
 {
@@ -34,13 +40,13 @@ void WebUIRequestHandler::handleRequest(HTTPServerRequest & request, HTTPServerR
     if (request.getURI().starts_with("/play"))
     {
         response.setStatusAndReason(Poco::Net::HTTPResponse::HTTP_OK);
-        *response.send() << getResource("play.html");
+        *response.send() << std::string_view(reinterpret_cast<const char *>(gresource_play_htmlData), gresource_play_htmlSize);
     }
     else if (request.getURI().starts_with("/dashboard"))
     {
         response.setStatusAndReason(Poco::Net::HTTPResponse::HTTP_OK);
 
-        std::string html(getResource("dashboard.html"));
+        std::string html(reinterpret_cast<const char *>(gresource_dashboard_htmlData), gresource_dashboard_htmlSize);
 
         /// Replace a link to external JavaScript file to embedded file.
         /// This allows to open the HTML without running a server and to host it on server.
@@ -55,7 +61,7 @@ void WebUIRequestHandler::handleRequest(HTTPServerRequest & request, HTTPServerR
     else if (request.getURI() == "/js/uplot.js")
     {
         response.setStatusAndReason(Poco::Net::HTTPResponse::HTTP_OK);
-        *response.send() << getResource("js/uplot.js");
+        *response.send() << std::string_view(reinterpret_cast<const char *>(gresource_uplot_jsData), gresource_uplot_jsSize);
     }
     else
     {

src/Storages/System/CMakeLists.txt

@@ -30,7 +30,6 @@ endif()
 add_dependencies(generate-source generate-contributors)
 
 set(GENERATED_LICENSES_SRC "${CMAKE_CURRENT_BINARY_DIR}/StorageSystemLicenses.generated.cpp")
-set(GENERATED_TIMEZONES_SRC "${CMAKE_CURRENT_BINARY_DIR}/StorageSystemTimeZones.generated.cpp")
 
 add_custom_command(
     OUTPUT StorageSystemLicenses.generated.cpp
@@ -38,23 +37,13 @@ add_custom_command(
     WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
 
 list (APPEND storages_system_sources ${GENERATED_LICENSES_SRC})
-list (APPEND storages_system_sources ${GENERATED_TIMEZONES_SRC})
 
 # Overlength strings
 set_source_files_properties(${GENERATED_LICENSES_SRC} PROPERTIES COMPILE_FLAGS -w)
 
-include(${ClickHouse_SOURCE_DIR}/cmake/embed_binary.cmake)
-clickhouse_embed_binaries(
-        TARGET information_schema_metadata
-        RESOURCE_DIR "${CMAKE_CURRENT_SOURCE_DIR}/InformationSchema/"
-        RESOURCES schemata.sql tables.sql views.sql columns.sql
-)
-
 list (SORT storages_system_sources) # Reproducible build
 add_library(clickhouse_storages_system ${storages_system_sources})
 
-add_dependencies(clickhouse_storages_system information_schema_metadata)
-
 target_link_libraries(clickhouse_storages_system PRIVATE
     dbms
     common
@@ -62,5 +51,6 @@ target_link_libraries(clickhouse_storages_system PRIVATE
     clickhouse_common_zookeeper
     clickhouse_parsers
     Poco::JSON
-    INTERFACE "-Wl,${WHOLE_ARCHIVE} $<TARGET_FILE:information_schema_metadata> -Wl,${NO_WHOLE_ARCHIVE}"
 )
+
+target_include_directories(clickhouse_storages_system PRIVATE InformationSchema)

src/Storages/System/attachInformationSchemaTables.cpp

@@ -3,14 +3,21 @@
 #include <Storages/System/attachSystemTablesImpl.h>
 #include <Parsers/ParserCreateQuery.h>
 #include <Parsers/parseQuery.h>
-#include <Common/getResource.h>
+#include <incbin.h>
+
+/// Embedded SQL definitions
+INCBIN(resource_schemata_sql, "schemata.sql");
+INCBIN(resource_tables_sql, "tables.sql");
+INCBIN(resource_views_sql, "views.sql");
+INCBIN(resource_columns_sql, "columns.sql");
+
 
 namespace DB
 {
 
 /// View structures are taken from http://www.contrib.andrew.cmu.edu/~shadow/sql/sql1992.txt
 
-static void createInformationSchemaView(ContextMutablePtr context, IDatabase & database, const String & view_name)
+static void createInformationSchemaView(ContextMutablePtr context, IDatabase & database, const String & view_name, std::string_view query)
 {
     try
     {
@@ -21,12 +28,11 @@ static void createInformationSchemaView(ContextMutablePtr context, IDatabase & d
         bool is_uppercase = database.getDatabaseName() == DatabaseCatalog::INFORMATION_SCHEMA_UPPERCASE;
 
         String metadata_resource_name = view_name + ".sql";
-        auto attach_query = getResource(metadata_resource_name);
+        if (query.empty())
-        if (attach_query.empty())
             return;
 
         ParserCreateQuery parser;
-        ASTPtr ast = parseQuery(parser, attach_query.data(), attach_query.data() + attach_query.size(),
+        ASTPtr ast = parseQuery(parser, query.data(), query.data() + query.size(),
                                 "Attach query from embedded resource " + metadata_resource_name,
                                 DBMS_DEFAULT_MAX_QUERY_SIZE, DBMS_DEFAULT_MAX_PARSER_DEPTH);
 
@@ -50,10 +56,10 @@ static void createInformationSchemaView(ContextMutablePtr context, IDatabase & d
 
 void attachInformationSchema(ContextMutablePtr context, IDatabase & information_schema_database)
 {
-    createInformationSchemaView(context, information_schema_database, "schemata");
+    createInformationSchemaView(context, information_schema_database, "schemata", std::string_view(reinterpret_cast<const char *>(gresource_schemata_sqlData), gresource_schemata_sqlSize));
-    createInformationSchemaView(context, information_schema_database, "tables");
+    createInformationSchemaView(context, information_schema_database, "tables", std::string_view(reinterpret_cast<const char *>(gresource_tables_sqlData), gresource_tables_sqlSize));
-    createInformationSchemaView(context, information_schema_database, "views");
+    createInformationSchemaView(context, information_schema_database, "views", std::string_view(reinterpret_cast<const char *>(gresource_views_sqlData), gresource_views_sqlSize));
-    createInformationSchemaView(context, information_schema_database, "columns");
+    createInformationSchemaView(context, information_schema_database, "columns", std::string_view(reinterpret_cast<const char *>(gresource_columns_sqlData), gresource_columns_sqlSize));
 }
 
 }

tests/queries/0_stateless/02415_all_new_functions_must_be_documented.reference

@@ -238,10 +238,6 @@ defaultValueOfArgumentType
 defaultValueOfTypeName
 degrees
 demangle
-detectCharset
-detectLanguageUnknown
-detectProgrammingLanguage
-detectTonality
 divide
 dotProduct
 dumpColumnStructure

tests/queries/0_stateless/02415_all_new_functions_must_be_documented.sql

@@ -15,5 +15,7 @@ AND name NOT IN (
     'h3ToGeoBoundary', 'h3ToParent', 'h3ToString', 'h3UnidirectionalEdgeIsValid', 'h3kRing', 'stringToH3',
     'geoToS2', 's2CapContains', 's2CapUnion', 's2CellsIntersect', 's2GetNeighbors', 's2RectAdd', 's2RectContains', 's2RectIntersection', 's2RectUnion', 's2ToGeo',
     'normalizeUTF8NFC', 'normalizeUTF8NFD', 'normalizeUTF8NFKC', 'normalizeUTF8NFKD',
-    'lemmatize', 'tokenize', 'stem', 'synonyms' -- these functions are not enabled in fast test
+    'lemmatize', 'tokenize', 'stem', 'synonyms',
+    'detectCharset', 'detectLanguageUnknown', 'detectProgrammingLanguage', 'detectTonality'
+     -- these functions are not enabled in fast test
 ) ORDER BY name;