ARROW-6312: [C++] Add support for "pkg-config --static arrow"

cpp/CMakeLists.txt

@@ -507,6 +507,11 @@ endif()
 include(BuildUtils)
 enable_testing()
 
+# For arrow.pc. Requires.private and Libs.private are used when
+# "pkg-config --libs --static arrow" is used.
+set(ARROW_PC_REQUIRES_PRIVATE)
+set(ARROW_PC_LIBS_PRIVATE)
+
 include(ThirdpartyToolchain)
 
 # Add common flags
@@ -855,8 +860,9 @@ endif()
 
 set(ARROW_SYSTEM_LINK_LIBS)
 
-if(THREADS_FOUND)
-  list(APPEND ARROW_SYSTEM_LINK_LIBS Threads::Threads)
+list(APPEND ARROW_SYSTEM_LINK_LIBS Threads::Threads)
+if(CMAKE_THREAD_LIBS_INIT)
+  string(APPEND ARROW_PC_LIBS_PRIVATE " ${CMAKE_THREAD_LIBS_INIT}")
 endif()
 
 if(WIN32)

cpp/cmake_modules/ThirdpartyToolchain.cmake

@@ -196,6 +196,7 @@ endmacro()
 macro(resolve_dependency DEPENDENCY_NAME)
   set(options)
   set(one_value_args HAVE_ALT IS_RUNTIME_DEPENDENCY REQUIRED_VERSION USE_CONFIG)
+  set(multi_value_args PC_PACKAGE_NAMES)
   cmake_parse_arguments(ARG
                         "${options}"
                         "${one_value_args}"
@@ -236,6 +237,13 @@ macro(resolve_dependency DEPENDENCY_NAME)
   if(${DEPENDENCY_NAME}_SOURCE STREQUAL "SYSTEM" AND ARG_IS_RUNTIME_DEPENDENCY)
     provide_find_module(${PACKAGE_NAME})
     list(APPEND ARROW_SYSTEM_DEPENDENCIES ${PACKAGE_NAME})
+    find_package(PkgConfig QUIET)
+    foreach(ARG_PC_PACKAGE_NAME ${ARG_PC_PACKAGE_NAMES})
+      pkg_check_modules(${ARG_PC_PACKAGE_NAME}_PC ${ARG_PC_PACKAGE_NAME} QUIET)
+      if(${${ARG_PC_PACKAGE_NAME}_PC_FOUND})
+        string(APPEND ARROW_PC_REQUIRES_PRIVATE " ${ARG_PC_PACKAGE_NAME}")
+      endif()
+    endforeach()
   endif()
 endmacro()
 
@@ -933,7 +941,11 @@ macro(build_snappy)
 endmacro()
 
 if(ARROW_WITH_SNAPPY)
-  resolve_dependency(Snappy)
+  resolve_dependency(Snappy PC_PACKAGE_NAMES snappy)
+  if(${Snappy_SOURCE} STREQUAL "SYSTEM" AND NOT snappy_PC_FOUND)
+    get_target_property(SNAPPY_LIB Snappy::snappy IMPORTED_LOCATION)
+    string(APPEND ARROW_PC_LIBS_PRIVATE " ${SNAPPY_LIB}")
+  endif()
   # TODO: Don't use global includes but rather target_include_directories
   get_target_property(SNAPPY_INCLUDE_DIRS Snappy::snappy INTERFACE_INCLUDE_DIRECTORIES)
   include_directories(SYSTEM ${SNAPPY_INCLUDE_DIRS})
@@ -998,7 +1010,7 @@ macro(build_brotli)
 endmacro()
 
 if(ARROW_WITH_BROTLI)
-  resolve_dependency(Brotli)
+  resolve_dependency(Brotli PC_PACKAGE_NAMES libbrotlidec libbrotlienc)
   # TODO: Don't use global includes but rather target_include_directories
   get_target_property(BROTLI_INCLUDE_DIR Brotli::brotlicommon
                       INTERFACE_INCLUDE_DIRECTORIES)
@@ -1079,9 +1091,9 @@ macro(build_glog)
   )
   set(GLOG_CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIC")
   set(GLOG_CMAKE_C_FLAGS "${EP_C_FLAGS} -fPIC")
-  if(Threads::Threads)
-    set(GLOG_CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fPIC -pthread")
-    set(GLOG_CMAKE_C_FLAGS "${EP_C_FLAGS} -fPIC -pthread")
+  if(CMAKE_THREAD_LIBS_INIT)
+    set(GLOG_CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${CMAKE_THREAD_LIBS_INIT}")
+    set(GLOG_CMAKE_C_FLAGS "${EP_C_FLAGS} ${CMAKE_THREAD_LIBS_INIT}")
   endif()
 
   if(APPLE)
@@ -1117,7 +1129,7 @@ macro(build_glog)
 endmacro()
 
 if(ARROW_USE_GLOG)
-  resolve_dependency(GLOG)
+  resolve_dependency(GLOG PC_PACKAGE_NAMES libglog)
   # TODO: Don't use global includes but rather target_include_directories
   get_target_property(GLOG_INCLUDE_DIR glog::glog INTERFACE_INCLUDE_DIRECTORIES)
   include_directories(SYSTEM ${GLOG_INCLUDE_DIR})
@@ -1292,7 +1304,11 @@ if(ARROW_WITH_THRIFT)
   # to build Boost, so don't look again if already found.
   if(NOT Thrift_FOUND AND NOT THRIFT_FOUND)
     # Thrift c++ code generated by 0.13 requires 0.11 or greater
-    resolve_dependency(Thrift REQUIRED_VERSION 0.11.0)
+    resolve_dependency(Thrift
+                       REQUIRED_VERSION
+                       0.11.0
+                       PC_PACKAGE_NAMES
+                       thrift)
   endif()
   # TODO: Don't use global includes but rather target_include_directories
   include_directories(SYSTEM ${THRIFT_INCLUDE_DIR})
@@ -1392,7 +1408,11 @@ if(ARROW_WITH_PROTOBUF)
   else()
     set(ARROW_PROTOBUF_REQUIRED_VERSION "2.6.1")
   endif()
-  resolve_dependency(Protobuf REQUIRED_VERSION ${ARROW_PROTOBUF_REQUIRED_VERSION})
+  resolve_dependency(Protobuf
+                     REQUIRED_VERSION
+                     ${ARROW_PROTOBUF_REQUIRED_VERSION}
+                     PC_PACKAGE_NAMES
+                     protobuf)
 
   if(ARROW_PROTOBUF_USE_SHARED AND MSVC_TOOLCHAIN)
     add_definitions(-DPROTOBUF_USE_DLLS)
@@ -1825,7 +1845,11 @@ if(ARROW_BUILD_BENCHMARKS)
     # ci/conda_env_cpp.yml.
     set(BENCHMARK_REQUIRED_VERSION 0.0.0)
   endif()
-  resolve_dependency(benchmark REQUIRED_VERSION ${BENCHMARK_REQUIRED_VERSION})
+  resolve_dependency(benchmark
+                     REQUIRED_VERSION
+                     ${BENCHMARK_REQUIRED_VERSION}
+                     IS_RUNTIME_DEPENDENCY
+                     FALSE)
   # TODO: Don't use global includes but rather target_include_directories
   get_target_property(BENCHMARK_INCLUDE_DIR benchmark::benchmark
                       INTERFACE_INCLUDE_DIRECTORIES)
@@ -1940,7 +1964,7 @@ macro(build_zlib)
 endmacro()
 
 if(ARROW_WITH_ZLIB)
-  resolve_dependency(ZLIB)
+  resolve_dependency(ZLIB PC_PACKAGE_NAMES zlib)
 
   # TODO: Don't use global includes but rather target_include_directories
   get_target_property(ZLIB_INCLUDE_DIR ZLIB::ZLIB INTERFACE_INCLUDE_DIRECTORIES)
@@ -1996,7 +2020,7 @@ macro(build_lz4)
 endmacro()
 
 if(ARROW_WITH_LZ4)
-  resolve_dependency(Lz4)
+  resolve_dependency(Lz4 PC_PACKAGE_NAMES liblz4)
 
   # TODO: Don't use global includes but rather target_include_directories
   get_target_property(LZ4_INCLUDE_DIR LZ4::lz4 INTERFACE_INCLUDE_DIRECTORIES)
@@ -2060,7 +2084,7 @@ macro(build_zstd)
 endmacro()
 
 if(ARROW_WITH_ZSTD)
-  resolve_dependency(zstd)
+  resolve_dependency(zstd PC_PACKAGE_NAMES libzstd)
 
   if(TARGET zstd::libzstd)
     set(ARROW_ZSTD_LIBZSTD zstd::libzstd)
@@ -2120,7 +2144,14 @@ macro(build_re2)
 endmacro()
 
 if(ARROW_WITH_RE2)
+  # Don't specify "PC_PACKAGE_NAMES re2" here because re2.pc may
+  # include -std=c++11. It's not compatible with C source and C++
+  # source not uses C++ 11.
   resolve_dependency(re2 HAVE_ALT TRUE)
+  if(${re2_SOURCE} STREQUAL "SYSTEM")
+    get_target_property(RE2_LIB re2::re2 IMPORTED_LOCATION)
+    string(APPEND ARROW_PC_LIBS_PRIVATE " ${RE2_LIB}")
+  endif()
   add_definitions(-DARROW_WITH_RE2)
 
   # TODO: Don't use global includes but rather target_include_directories
@@ -2169,6 +2200,9 @@ endmacro()
 
 if(ARROW_WITH_BZ2)
   resolve_dependency(BZip2)
+  if(${BZip2_SOURCE} STREQUAL "SYSTEM")
+    string(APPEND ARROW_PC_LIBS_PRIVATE " ${BZIP2_LIBRARIES}")
+  endif()
 
   if(NOT TARGET BZip2::BZip2)
     add_library(BZip2::BZip2 UNKNOWN IMPORTED)
@@ -2219,7 +2253,11 @@ macro(build_utf8proc)
 endmacro()
 
 if(ARROW_WITH_UTF8PROC)
-  resolve_dependency(utf8proc REQUIRED_VERSION "2.2.0")
+  resolve_dependency(utf8proc
+                     REQUIRED_VERSION
+                     "2.2.0"
+                     PC_PACKAGE_NAMES
+                     libutf8proc)
 
   add_definitions(-DARROW_WITH_UTF8PROC)
 
@@ -2287,7 +2325,11 @@ endmacro()
 # Dependencies for Arrow Flight RPC
 
 macro(build_grpc)
-  resolve_dependency(c-ares HAVE_ALT TRUE)
+  resolve_dependency(c-ares
+                     HAVE_ALT
+                     TRUE
+                     PC_PACKAGE_NAMES
+                     libcares)
   # TODO: Don't use global includes but rather target_include_directories
   get_target_property(c-ares_INCLUDE_DIR c-ares::cares INTERFACE_INCLUDE_DIRECTORIES)
   include_directories(SYSTEM ${c-ares_INCLUDE_DIR})
@@ -2551,7 +2593,9 @@ if(ARROW_WITH_GRPC)
                      HAVE_ALT
                      TRUE
                      REQUIRED_VERSION
-                     ${ARROW_GRPC_REQUIRED_VERSION})
+                     ${ARROW_GRPC_REQUIRED_VERSION}
+                     PC_PACKAGE_NAMES
+                     grpc++)
 
   # TODO: Don't use global includes but rather target_include_directories
   get_target_property(GRPC_INCLUDE_DIR gRPC::grpc++ INTERFACE_INCLUDE_DIRECTORIES)

cpp/examples/minimal_build/minimal.dockerfile

@@ -22,5 +22,6 @@ ENV DEBIAN_FRONTEND=noninteractive
 RUN apt-get update -y -q && \
     apt-get install -y -q --no-install-recommends \
       build-essential \
-      cmake && \
+      cmake \
+      pkg-config && \
     apt-get clean && rm -rf /var/lib/apt/lists*

cpp/examples/minimal_build/run_static.sh

@@ -67,10 +67,12 @@ popd
 
 echo
 echo "=="
+echo "== CMake:"
 echo "== Building example project using Arrow C++ library"
 echo "=="
 echo
 
+rm -rf $EXAMPLE_BUILD_DIR
 mkdir -p $EXAMPLE_BUILD_DIR
 pushd $EXAMPLE_BUILD_DIR
 
@@ -81,10 +83,39 @@ popd
 
 echo
 echo "=="
+echo "== CMake:"
 echo "== Running example project"
 echo "=="
 echo
 
 pushd $EXAMPLE_DIR
 
-${EXAMPLE_BUILD_DIR}/arrow_example
+$EXAMPLE_BUILD_DIR/arrow_example
+
+echo
+echo "=="
+echo "== pkg-config"
+echo "== Building example project using Arrow C++ library"
+echo "=="
+echo
+
+rm -rf $EXAMPLE_BUILD_DIR
+mkdir -p $EXAMPLE_BUILD_DIR
+${CXX:-c++} \
+  -o $EXAMPLE_BUILD_DIR/arrow_example \
+  $EXAMPLE_DIR/example.cc \
+  $(PKG_CONFIG_PATH=$ARROW_BUILD_DIR/lib/pkgconfig \
+     pkg-config --cflags --libs --static arrow)
+
+popd
+
+echo
+echo "=="
+echo "== pkg-config:"
+echo "== Running example project"
+echo "=="
+echo
+
+pushd $EXAMPLE_DIR
+
+$EXAMPLE_BUILD_DIR/arrow_example

cpp/examples/minimal_build/system_dependency.dockerfile

@@ -37,6 +37,7 @@ RUN apt-get update -y -q && \
       libthrift-dev \
       libutf8proc-dev \
       libzstd-dev \
+      pkg-config \
       protobuf-compiler \
       rapidjson-dev \
       zlib1g-dev && \

cpp/src/arrow/CMakeLists.txt

@@ -489,6 +489,21 @@ endif()
 
 set(ARROW_ALL_SRCS ${ARROW_SRCS} ${ARROW_C_SRCS})
 
+if(ARROW_BUILD_STATIC AND ARROW_BUNDLED_STATIC_LIBS)
+  set(ARROW_BUILD_BUNDLED_DEPENDENCIES TRUE)
+else()
+  set(ARROW_BUILD_BUNDLED_DEPENDENCIES FALSE)
+endif()
+
+if(ARROW_BUILD_BUNDLED_DEPENDENCIES)
+  string(APPEND ARROW_PC_LIBS_PRIVATE " -larrow_bundled_dependencies")
+endif()
+# Need -latomic on Raspbian.
+# See also: https://issues.apache.org/jira/browse/ARROW-12860
+if(${CMAKE_SYSTEM_NAME} STREQUAL "Linux" AND ${CMAKE_SYSTEM_PROCESSOR} MATCHES "armv7")
+  string(APPEND ARROW_PC_LIBS_PRIVATE " -latomic")
+endif()
+
 add_arrow_lib(arrow
               CMAKE_PACKAGE_NAME
               Arrow
@@ -535,7 +550,7 @@ if(ARROW_WITH_BACKTRACE)
   endforeach()
 endif()
 
-if(ARROW_BUILD_STATIC AND ARROW_BUNDLED_STATIC_LIBS)
+if(ARROW_BUILD_BUNDLED_DEPENDENCIES)
   arrow_car(_FIRST_LIB ${ARROW_BUNDLED_STATIC_LIBS})
   arrow_cdr(_OTHER_LIBS ${ARROW_BUNDLED_STATIC_LIBS})
   create_merged_static_lib(arrow_bundled_dependencies

cpp/src/arrow/arrow.pc.in

@@ -25,5 +25,7 @@ full_so_version=@ARROW_FULL_SO_VERSION@
 Name: Apache Arrow
 Description: Arrow is a set of technologies that enable big-data systems to process and move data fast.
 Version: @ARROW_VERSION@
+Requires.private:@ARROW_PC_REQUIRES_PRIVATE@
 Libs: -L${libdir} -larrow
+Libs.private:@ARROW_PC_LIBS_PRIVATE@
 Cflags: -I${includedir}