GH-1408: Add bundled utf8_range headers to the macOS Gandiva JNI build

.github/workflows/dev.yml

@@ -42,7 +42,7 @@ jobs:
 
       - uses: actions/setup-python@v7
         with:
-          python-version: '3.x'
+          python-version: '3.14'
       - name: pre-commit (cache)
         uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
         with:

CMakeLists.txt

@@ -41,6 +41,10 @@ set(ARROW_JAVA_JNI_ABSL_INCLUDE_DIR
     ""
     CACHE PATH "Path to Abseil source headers for the Gandiva JNI build")
 mark_as_advanced(ARROW_JAVA_JNI_ABSL_INCLUDE_DIR)
+set(ARROW_JAVA_JNI_UTF8_RANGE_INCLUDE_DIR
+    ""
+    CACHE PATH "Path to utf8_range source headers for the Gandiva JNI build")
+mark_as_advanced(ARROW_JAVA_JNI_UTF8_RANGE_INCLUDE_DIR)
 
 include(GNUInstallDirs)
 

ci/docker/vcpkg-jni.dockerfile

@@ -21,7 +21,15 @@ FROM ${base}
 # Install Java
 # We need Java for JNI headers, but we don't invoke Maven in this build.
 ARG java=17
-RUN dnf install -y java-$java-openjdk-devel && dnf clean all
+# The repository metadata is refreshed and the install retried because the
+# baseos mirror intermittently returns an empty repomd.xml.
+RUN dnf clean all && rm -rf /var/cache/dnf && \
+    for i in 1 2 3; do \
+      dnf install -y java-$java-openjdk-devel && break; \
+      if [ "$i" = 3 ]; then exit 1; fi; \
+      sleep 15; \
+    done && \
+    dnf clean all
 
 # For ci/scripts/{cpp,java}_*.sh
 ENV ARROW_HOME=/tmp/local \

ci/scripts/jni_macos_build.sh

@@ -86,9 +86,17 @@ if [ ! -d "${absl_include_dir}/absl" ]; then
   exit 1
 fi
 
+# Protobuf 33+ headers include utf8_validity.h from the bundled utf8_range.
+utf8_range_include_dir="${build_dir}/cpp/_deps/protobuf-src/third_party/utf8_range"
+if [ ! -f "${utf8_range_include_dir}/utf8_validity.h" ]; then
+  echo "Bundled utf8_range headers were not found in ${utf8_range_include_dir}" >&2
+  exit 1
+fi
+
 JAVA_JNI_CMAKE_ARGS="-DProtobuf_ROOT=${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+=" -DARROW_JAVA_JNI_UTF8_RANGE_INCLUDE_DIR=${utf8_range_include_dir}"
 export JAVA_JNI_CMAKE_ARGS
 "${source_dir}/ci/scripts/jni_build.sh" \
   "${source_dir}" \

gandiva/CMakeLists.txt

@@ -58,6 +58,10 @@ if(ARROW_JAVA_JNI_ABSL_INCLUDE_DIR)
   target_include_directories(arrow_java_jni_gandiva SYSTEM
                              PRIVATE "${ARROW_JAVA_JNI_ABSL_INCLUDE_DIR}")
 endif()
+if(ARROW_JAVA_JNI_UTF8_RANGE_INCLUDE_DIR)
+  target_include_directories(arrow_java_jni_gandiva SYSTEM
+                             PRIVATE "${ARROW_JAVA_JNI_UTF8_RANGE_INCLUDE_DIR}")
+endif()
 target_link_libraries(arrow_java_jni_gandiva
                       arrow_java_jni_gandiva_headers
                       jni