apacheGH-40855: [C++][ORC] Don't use std::filesystem with ORC 2.0.0 o…

…r later

If we use `std::filesystem`, we need `-lstdc++fs` with GCC 8 and
`-lc++fs` for clang 7. We don't want to maintain CMake code for GCC
8/clang 7. So this avoids using `std::filesystem` with ORC 2.0.0 or
later.

cpp/cmake_modules/FindorcAlt.cmake

@@ -29,6 +29,7 @@ endif()
 find_package(orc ${find_package_args})
 if(orc_FOUND)
   set(orcAlt_FOUND TRUE)
+  set(orcAlt_VERSION ${orc_VERSION})
   return()
 endif()
 
@@ -51,8 +52,17 @@ else()
             NAMES orc/orc-config.hh
             PATH_SUFFIXES ${ARROW_INCLUDE_PATH_SUFFIXES})
 endif()
+if(ORC_INCLUDE_DIR)
+  file(READ "${ORC_INCLUDE_DIR}/orc/orc-config.hh" ORC_CONFIG_HH_CONTENT)
+  string(REGEX MATCH "#define ORC_VERSION \"[0-9.]+\"" ORC_VERSION_DEFINITION
+               "${ORC_CONFIG_HH_CONTENT}")
+  string(REGEX MATCH "[0-9.]+" ORC_VERSION "${ORC_VERSION_DEFINITION}")
+endif()
 
-find_package_handle_standard_args(orcAlt REQUIRED_VARS ORC_STATIC_LIB ORC_INCLUDE_DIR)
+find_package_handle_standard_args(
+  orcAlt
+  REQUIRED_VARS ORC_STATIC_LIB ORC_INCLUDE_DIR
+  VERSION_VAR ORC_VERSION)
 
 if(orcAlt_FOUND)
   if(NOT TARGET orc::orc)

cpp/cmake_modules/ThirdpartyToolchain.cmake

@@ -4559,6 +4559,15 @@ macro(build_orc)
     endif()
     target_link_libraries(orc::orc INTERFACE ${CMAKE_DL_LIBS})
   endif()
+  if(CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
+    if(CMAKE_CXX_COMPILER_VERSION VERSION_LESS "9")
+      target_link_libraries(orc::orc INTERFACE stdc++fs)
+    endif()
+  elseif(CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
+    if(CMAKE_CXX_COMPILER_VERSION VERSION_LESS "8")
+      target_link_libraries(orc::orc INTERFACE c++fs)
+    endif()
+  endif()
 
   add_dependencies(orc::orc orc_ep)
 
@@ -4568,6 +4577,11 @@ endmacro()
 if(ARROW_ORC)
   resolve_dependency(orc HAVE_ALT TRUE)
   target_link_libraries(orc::orc INTERFACE ${ARROW_PROTOBUF_LIBPROTOBUF})
+  if(ORC_VENDORED)
+    set(ARROW_ORC_VERSION ${ARROW_ORC_BUILD_VERSION})
+  else()
+    set(ARROW_ORC_VERSION ${orcAlt_VERSION})
+  endif()
   message(STATUS "Found ORC static library: ${ORC_STATIC_LIB}")
   message(STATUS "Found ORC headers: ${ORC_INCLUDE_DIR}")
 endif()

cpp/src/arrow/CMakeLists.txt

@@ -899,6 +899,10 @@ if(ARROW_ORC)
                            adapters/orc/util.cc)
   foreach(ARROW_ORC_TARGET ${ARROW_ORC_TARGETS})
     target_link_libraries(${ARROW_ORC_TARGET} PRIVATE orc::orc)
+    if(ARROW_ORC_VERSION VERSION_LESS "2.0.0")
+      target_compile_definitions(${ARROW_ORC_TARGET}
+                                 PRIVATE ARROW_ORC_NEED_TIME_ZONE_DATABASE_CHECK)
+    endif()
   endforeach()
 else()
   set(ARROW_ORC_TARGET_SHARED)

cpp/src/arrow/adapters/orc/adapter.cc

@@ -18,13 +18,16 @@
 #include "arrow/adapters/orc/adapter.h"
 
 #include <algorithm>
-#include <filesystem>
 #include <list>
 #include <memory>
 #include <sstream>
 #include <string>
 #include <vector>
 
+#ifdef ARROW_ORC_NEED_TIME_ZONE_DATABASE_CHECK
+#include <filesystem>
+#endif
+
 #include "arrow/adapters/orc/util.h"
 #include "arrow/builder.h"
 #include "arrow/io/interfaces.h"
@@ -183,11 +186,9 @@ liborc::RowReaderOptions DefaultRowReaderOptions() {
   return options;
 }
 
+#ifdef ARROW_ORC_NEED_TIME_ZONE_DATABASE_CHECK
 // Proactively check timezone database availability for ORC versions older than 2.0.0
 Status CheckTimeZoneDatabaseAvailability() {
-  if (GetOrcMajorVersion() >= 2) {
-    return Status::OK();
-  }
   auto tz_dir = std::getenv("TZDIR");
   bool is_tzdb_avaiable = tz_dir != nullptr
                               ? std::filesystem::exists(tz_dir)
@@ -200,6 +201,7 @@ Status CheckTimeZoneDatabaseAvailability() {
   }
   return Status::OK();
 }
+#endif
 
 }  // namespace
 
@@ -559,7 +561,9 @@ ORCFileReader::~ORCFileReader() {}
 
 Result<std::unique_ptr<ORCFileReader>> ORCFileReader::Open(
     const std::shared_ptr<io::RandomAccessFile>& file, MemoryPool* pool) {
+#ifdef ARROW_ORC_NEED_TIME_ZONE_DATABASE_CHECK
   RETURN_NOT_OK(CheckTimeZoneDatabaseAvailability());
+#endif
   auto result = std::unique_ptr<ORCFileReader>(new ORCFileReader());
   RETURN_NOT_OK(result->impl_->Open(file, pool));
   return std::move(result);
@@ -826,7 +830,9 @@ ORCFileWriter::ORCFileWriter() { impl_.reset(new ORCFileWriter::Impl()); }
 
 Result<std::unique_ptr<ORCFileWriter>> ORCFileWriter::Open(
     io::OutputStream* output_stream, const WriteOptions& writer_options) {
+#ifdef ARROW_ORC_NEED_TIME_ZONE_DATABASE_CHECK
   RETURN_NOT_OK(CheckTimeZoneDatabaseAvailability());
+#endif
   std::unique_ptr<ORCFileWriter> result =
       std::unique_ptr<ORCFileWriter>(new ORCFileWriter());
   Status status = result->impl_->Open(output_stream, writer_options);