DO NOT MERGE: Fix JNI builds

.github/workflows/rc.yml

@@ -197,20 +197,13 @@ jobs:
           python-version: 3.12
       - name: Install Archery
         run: pip install -e arrow/dev/archery[all]
+      - name: Set up Java
+        uses: actions/setup-java@v6
+        with:
+          java-version: '17'
+          distribution: 'temurin'
       - name: Install dependencies
         run: |
-          # We want to use llvm@14 to avoid shared z3
-          # dependency. llvm@14 doesn't depend on z3 and llvm depends
-          # on z3. And Homebrew's z3 provides only shared library. It
-          # doesn't provides static z3 because z3's CMake doesn't accept
-          # building both shared and static libraries at once.
-          # See also: Z3_BUILD_LIBZ3_SHARED in
-          # https://github.com/Z3Prover/z3/blob/master/README-CMake.md
-          #
-          # If llvm is installed, Apache Arrow C++ uses llvm rather than
-          # llvm@14 because llvm is newer than llvm@14.
-          brew uninstall llvm || :
-
           # We can remove this when we drop support for
           # macos-15-intel. because macos-14 or later with arm64 uses /opt/homebrew/
           # not /usr/local/.
@@ -228,6 +221,9 @@ jobs:
           done
           brew install --overwrite python3
 
+          # Use a version supported by Gandiva with prebuilt macOS bottles.
+          brew install --force-bottle llvm@18
+
           if [ "$(uname -m)" = "arm64" ]; then
             # pkg-config formula is deprecated but it's still installed
             # in GitHub Actions runner now. We can remove this once
@@ -236,7 +232,7 @@ jobs:
             brew uninstall pkg-config@0.29.2 || :
           fi
 
-          # We don't use Homebrew's aws-sdk-cpp and gRPC. See the
+          # We don't use Homebrew's aws-sdk-cpp, gRPC, RE2 or Protobuf. See the
           # "brew uninstall" calls below for details. So we don't install
           # them instead of installing and uninstalling them.
           #
@@ -251,8 +247,10 @@ jobs:
           #
           # We don't use Homebrew's Node.js too. Homebrew doesn't provide
           # a bottle for it on x86_64 macOS for the same reason.
-          HOMEBREW_BUNDLE_BREW_SKIP="aws-sdk-cpp grpc node" \
-            brew bundle --file=arrow/cpp/Brewfile
+          # LLVM is installed explicitly above instead of using the latest version.
+          # Keep installed dependencies instead of rebuilding newer releases.
+          HOMEBREW_BUNDLE_BREW_SKIP="aws-sdk-cpp grpc re2 protobuf node llvm" \
+            brew bundle --no-upgrade --verbose --file=arrow/cpp/Brewfile
           # We want to link aws-sdk-cpp statically but Homebrew's
           # aws-sdk-cpp provides only shared library. If we have
           # Homebrew's aws-sdk-cpp, our build mix Homebrew's
@@ -267,14 +265,15 @@ jobs:
           # We uninstall Homebrew's RE2 to ensure using bundled RE2.
           brew uninstall grpc || : # gRPC depends on RE2
           brew uninstall grpc@1.54 || : # gRPC 1.54 may be installed too
-          brew uninstall re2
           # We want to use bundled Protobuf for static linking. If
           # Homebrew's Protobuf is installed, its library file may be
           # used on test  We uninstall Homebrew's Protobuf to ensure using
           # bundled Protobuf.
-          brew uninstall protobuf
-
-          brew bundle --file=Brewfile
+          for package in re2 protobuf; do
+            if brew list --versions "${package}" >/dev/null; then
+              brew uninstall "${package}"
+            fi
+          done
       - name: Prepare ccache
         run: |
           echo "CCACHE_DIR=${PWD}/ccache" >> ${GITHUB_ENV}
@@ -287,8 +286,6 @@ jobs:
       - name: Build
         run: |
           set -e
-          # make brew Java available to CMake
-          export JAVA_HOME=$(brew --prefix openjdk@17)/libexec/openjdk.jdk/Contents/Home
           ci/scripts/jni_macos_build.sh . arrow build jni
       - name: Compress into single artifact to keep directory structure
         run: tar -cvzf jni-macos-${{ matrix.platform.arch }}.tar.gz jni/

Brewfile

@@ -16,4 +16,3 @@
 # under the License.
 
 brew "openjdk@17"
-brew "sccache"

ci/scripts/jni_macos_build.sh

@@ -69,12 +69,22 @@ export ARROW_PARQUET=ON
 
 export AWS_EC2_METADATA_DISABLED=TRUE
 
+llvm_root="$(brew --prefix llvm@18)"
+"${llvm_root}/bin/llvm-config" --version
+llvm_cmake_args=(
+  "-DLLVM_ROOT=${llvm_root}"
+  "-DLLVM_DIR=${llvm_root}/lib/cmake/llvm"
+  "-DCLANG_EXECUTABLE=${llvm_root}/bin/clang"
+  "-DLLVM_LINK_EXECUTABLE=${llvm_root}/bin/llvm-link"
+)
+
 cmake \
   -S "${arrow_dir}/cpp" \
   -B "${build_dir}/cpp" \
   --preset=ninja-release-jni-macos \
   -Dabsl_SOURCE=BUNDLED \
   -Dsimdjson_SOURCE=BUNDLED \
+  "${llvm_cmake_args[@]}" \
   -DCMAKE_INSTALL_PREFIX="${install_dir}"
 cmake --build "${build_dir}/cpp" --target install
 github_actions_group_end
@@ -88,6 +98,7 @@ fi
 JAVA_JNI_CMAKE_ARGS="-DProtobuf_ROOT=${build_dir}/cpp/_deps/protobuf-build"
 JAVA_JNI_CMAKE_ARGS+=" -DProtobuf_SRC_ROOT_FOLDER=${build_dir}/cpp/_deps/protobuf-src"
 JAVA_JNI_CMAKE_ARGS+=" -DARROW_JAVA_JNI_ABSL_INCLUDE_DIR=${absl_include_dir}"
+JAVA_JNI_CMAKE_ARGS+=" ${llvm_cmake_args[*]}"
 export JAVA_JNI_CMAKE_ARGS
 "${source_dir}/ci/scripts/jni_build.sh" \
   "${source_dir}" \

dataset/CMakeLists.txt

@@ -17,6 +17,7 @@
 
 find_package(ArrowDataset REQUIRED)
 find_package(ArrowSubstrait REQUIRED)
+find_package(ArrowS3 QUIET)
 
 include_directories(${CMAKE_CURRENT_BINARY_DIR} ${CMAKE_CURRENT_SOURCE_DIR}
                     ${JNI_INCLUDE_DIRS} ${JNI_HEADERS_DIR})
@@ -40,6 +41,11 @@ target_link_libraries(arrow_java_jni_dataset
                       ArrowDataset::arrow_dataset_static
                       ArrowSubstrait::arrow_substrait_static)
 
+# Older Arrow releases include S3 in arrow_static instead of a separate library.
+if(TARGET ArrowS3::arrow_s3_static)
+  target_link_libraries(arrow_java_jni_dataset ArrowS3::arrow_s3_static)
+endif()
+
 if(BUILD_TESTING)
   add_executable(arrow-java-jni-dataset-test src/main/cpp/jni_util_test.cc
                                              src/main/cpp/jni_util.cc)