Move all url functions to clickhouse_functions_url lib (#5696)

CMakeLists.txt

@@ -344,6 +344,7 @@ find_contrib_lib(metrohash)
 find_contrib_lib(btrie)
 find_contrib_lib(double-conversion)
 include (cmake/find_parquet.cmake)
+include (cmake/find_gperf.cmake)
 
 if (ENABLE_TESTS)
     include (cmake/find_gtest.cmake)

dbms/src/AggregateFunctions/CMakeLists.txt

@@ -21,8 +21,8 @@ list(REMOVE_ITEM clickhouse_aggregate_functions_headers
 
 add_library(clickhouse_aggregate_functions ${clickhouse_aggregate_functions_sources})
 target_link_libraries(clickhouse_aggregate_functions PRIVATE dbms PUBLIC ${CITYHASH_LIBRARIES})
-target_include_directories (clickhouse_aggregate_functions BEFORE PRIVATE ${COMMON_INCLUDE_DIR})
+target_include_directories(clickhouse_aggregate_functions PRIVATE ${COMMON_INCLUDE_DIR})
 
-if (ENABLE_TESTS)
-    add_subdirectory (tests)
-endif ()
+if(ENABLE_TESTS)
+    add_subdirectory(tests)
+endif()

dbms/src/Functions/CMakeLists.txt

@@ -1,20 +1,9 @@
-include (${ClickHouse_SOURCE_DIR}/cmake/find_gperf.cmake)
 configure_file(config_functions.h.in ${CMAKE_CURRENT_BINARY_DIR}/include/config_functions.h)
 
 include(${ClickHouse_SOURCE_DIR}/cmake/dbms_glob_sources.cmake)
-
 add_headers_and_sources(clickhouse_functions ./GatherUtils)
 add_headers_and_sources(clickhouse_functions .)
 
-if (USE_GPERF)
-    add_custom_command(
-	    OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/tldLookup.cpp
-	    COMMAND ${GPERF} ${CMAKE_CURRENT_SOURCE_DIR}/tldLookup.gperf --output-file=${CMAKE_CURRENT_BINARY_DIR}/tldLookup.cpp
-    )
-
-    list(APPEND clickhouse_functions_sources ${CMAKE_CURRENT_BINARY_DIR}/tldLookup.cpp)
-endif ()
-
 list(REMOVE_ITEM clickhouse_functions_sources IFunction.cpp FunctionFactory.cpp FunctionHelpers.cpp)
 list(REMOVE_ITEM clickhouse_functions_headers IFunction.h FunctionFactory.h FunctionHelpers.h)
 
@@ -89,3 +78,6 @@ endif()
 if(USE_RAPIDJSON)
     target_include_directories(clickhouse_functions SYSTEM PRIVATE ${RAPIDJSON_INCLUDE_DIR})
 endif()
+
+add_subdirectory(URL)
+target_link_libraries(clickhouse_functions PRIVATE clickhouse_functions_url)

dbms/src/Functions/URL/CMakeLists.txt

@@ -0,0 +1,27 @@
+configure_file(config_functions_url.h.in ${CMAKE_CURRENT_BINARY_DIR}/include/config_functions_url.h)
+include(${ClickHouse_SOURCE_DIR}/cmake/dbms_glob_sources.cmake)
+add_headers_and_sources(clickhouse_functions_url .)
+
+if (USE_GPERF)
+    add_custom_command(
+	    OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/tldLookup.cpp
+	    COMMAND ${GPERF} ${CMAKE_CURRENT_SOURCE_DIR}/tldLookup.gperf --output-file=${CMAKE_CURRENT_BINARY_DIR}/tldLookup.cpp
+    )
+    list(APPEND clickhouse_functions_url_sources ${CMAKE_CURRENT_BINARY_DIR}/tldLookup.cpp)
+endif ()
+
+add_library(clickhouse_functions_url ${clickhouse_functions_url_sources} ${clickhouse_functions_url_headers})
+
+target_link_libraries(clickhouse_functions_url PRIVATE clickhouse_common_io)
+target_include_directories(clickhouse_functions_url PRIVATE ${CMAKE_CURRENT_BINARY_DIR}/include ${CMAKE_CURRENT_BINARY_DIR}/../include)
+
+if (CMAKE_BUILD_TYPE_UC STREQUAL "RELEASE" OR CMAKE_BUILD_TYPE_UC STREQUAL "RELWITHDEBINFO" OR CMAKE_BUILD_TYPE_UC STREQUAL "MINSIZEREL")
+    # Won't generate debug info for files with heavy template instantiation to achieve faster linking and lower size.
+    target_compile_options(clickhouse_functions_url PRIVATE "-g0")
+endif ()
+
+# TODO: move Functions/Regexps.h to some lib and use here
+if(USE_HYPERSCAN)
+    target_link_libraries(clickhouse_functions_url PRIVATE ${HYPERSCAN_LIBRARY})
+    target_include_directories(clickhouse_functions_url SYSTEM PRIVATE ${HYPERSCAN_INCLUDE_DIR})
+endif()

dbms/src/Functions/URL/URLHierarchy.cpp

@@ -1,6 +1,5 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsStringArray.h>
-#include <Functions/FunctionsURL.h>
 
 namespace DB
 {

dbms/src/Functions/URL/URLPathHierarchy.cpp

@@ -1,6 +1,5 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsStringArray.h>
-#include <Functions/FunctionsURL.h>
 
 namespace DB
 {

dbms/src/Functions/URL/basename.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/FunctionsURL.h>
+#include "FunctionsURL.h"
 #include <common/find_symbols.h>
 
 namespace DB

dbms/src/Functions/URL/config_functions_url.h.in

@@ -0,0 +1,5 @@
+#pragma once
+
+// .h autogenerated by cmake!
+
+#cmakedefine01 USE_GPERF

dbms/src/Functions/URL/cutFragment.cpp

@@ -1,5 +1,5 @@
 #include <Functions/FunctionFactory.h>
-#include <Functions/fragment.h>
+#include "fragment.h"
 #include <Functions/FunctionStringToString.h>
 
 namespace DB

dbms/src/Functions/URL/cutQueryString.cpp

@@ -1,5 +1,5 @@
 #include <Functions/FunctionFactory.h>
-#include <Functions/queryString.h>
+#include "queryString.h"
 #include <Functions/FunctionStringToString.h>
 
 namespace DB

dbms/src/Functions/URL/cutQueryStringAndFragment.cpp

@@ -1,5 +1,5 @@
 #include <Functions/FunctionFactory.h>
-#include <Functions/queryStringAndFragment.h>
+#include "queryStringAndFragment.h"
 #include <Functions/FunctionStringToString.h>
 
 namespace DB

dbms/src/Functions/URL/cutToFirstSignificantSubdomain.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/firstSignificantSubdomain.h>
+#include "firstSignificantSubdomain.h"
 
 
 namespace DB

dbms/src/Functions/URL/cutURLParameter.cpp

@@ -1,6 +1,5 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsStringSearch.h>
-#include <Functions/FunctionsURL.h>
 #include <common/find_symbols.h>
 
 namespace DB

dbms/src/Functions/URL/cutWWW.cpp

@@ -1,7 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/FunctionsURL.h>
-#include <Functions/protocol.h>
+#include "protocol.h"
 #include <common/find_symbols.h>
 
 

dbms/src/Functions/URL/decodeURLComponent.cpp

@@ -1,6 +1,5 @@
 #include <Common/hex.h>
 #include <Functions/FunctionFactory.h>
-#include <Functions/FunctionsURL.h>
 #include <Functions/FunctionStringToString.h>
 #include <common/find_symbols.h>
 

dbms/src/Functions/URL/domain.cpp

@@ -1,6 +1,7 @@
+#include "domain.h"
+
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/domain.h>
 
 namespace DB
 {

dbms/src/Functions/URL/domain.h

@@ -1,6 +1,6 @@
 #pragma once
 
-#include <Functions/protocol.h>
+#include "protocol.h"
 #include <common/find_symbols.h>
 #include <cstring>
 

dbms/src/Functions/URL/domainWithoutWWW.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/domain.h>
+#include "domain.h"
 
 namespace DB
 {

dbms/src/Functions/URL/extractURLParameter.cpp

@@ -1,6 +1,5 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsStringSearch.h>
-#include <Functions/FunctionsURL.h>
 #include <common/find_symbols.h>
 
 namespace DB

dbms/src/Functions/URL/extractURLParameterNames.cpp

@@ -1,6 +1,5 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsStringArray.h>
-#include <Functions/FunctionsURL.h>
 
 namespace DB
 {

dbms/src/Functions/URL/extractURLParameters.cpp

@@ -1,6 +1,5 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionsStringArray.h>
-#include <Functions/FunctionsURL.h>
 
 namespace DB
 {

dbms/src/Functions/URL/firstSignificantSubdomain.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/firstSignificantSubdomain.h>
+#include "firstSignificantSubdomain.h"
 
 
 namespace DB

dbms/src/Functions/URL/fragment.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/fragment.h>
+#include "fragment.h"
 
 namespace DB
 {

dbms/src/Functions/URL/fragment.h

@@ -1,6 +1,6 @@
 #pragma once
 
-#include <Functions/FunctionsURL.h>
+#include "FunctionsURL.h"
 #include <common/find_symbols.h>
 
 namespace DB

dbms/src/Functions/URL/path.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/FunctionsURL.h>
+#include "FunctionsURL.h"
 #include <common/find_symbols.h>
 
 namespace DB

dbms/src/Functions/URL/pathFull.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/FunctionsURL.h>
+#include "FunctionsURL.h"
 #include <common/find_symbols.h>
 
 namespace DB

dbms/src/Functions/URL/protocol.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/protocol.h>
+#include "protocol.h"
 
 
 namespace DB

dbms/src/Functions/URL/protocol.h

@@ -1,6 +1,6 @@
 #pragma once
 
-#include <Functions/FunctionsURL.h>
+#include "FunctionsURL.h"
 #include <Common/StringUtils/StringUtils.h>
 
 

dbms/src/Functions/URL/queryString.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/queryString.h>
+#include "queryString.h"
 
 namespace DB
 {

dbms/src/Functions/URL/queryString.h

@@ -1,6 +1,6 @@
 #pragma once
 
-#include <Functions/FunctionsURL.h>
+#include "FunctionsURL.h"
 #include <common/find_symbols.h>
 
 

dbms/src/Functions/URL/queryStringAndFragment.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/queryStringAndFragment.h>
+#include "queryStringAndFragment.h"
 
 namespace DB
 {

dbms/src/Functions/URL/queryStringAndFragment.h

@@ -1,6 +1,6 @@
 #pragma once
 
-#include <Functions/FunctionsURL.h>
+#include "FunctionsURL.h"
 #include <common/find_symbols.h>
 
 

dbms/src/Functions/URL/tldLookup.h

@@ -1,6 +1,6 @@
 #pragma once
 
-#include "config_functions.h"
+#include "config_functions_url.h"
 #if USE_GPERF
 // Definition of the class generated by gperf, present on gperf/tldLookup.gperf
 class tldLookupHash

dbms/src/Functions/URL/topLevelDomain.cpp

@@ -1,6 +1,6 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/FunctionStringToString.h>
-#include <Functions/domain.h>
+#include "domain.h"
 
 namespace DB
 {

dbms/src/Functions/config_functions.h.in

@@ -8,4 +8,3 @@
 #cmakedefine01 USE_HYPERSCAN
 #cmakedefine01 USE_SIMDJSON
 #cmakedefine01 USE_RAPIDJSON
-#cmakedefine01 USE_GPERF