GH-40507: [C++][ORC] Upgrade ORC to 2.0.0 (#40508)

wgtmac · web-flow · commit 4f39e6eac9f2 · 2024-03-28T18:17:33.000+01:00
### Rationale for this change This PR aims to upgrade to a new major version of Apache ORC: https://orc.apache.org/news/2024/03/08/ORC-2.0.0/ ### What changes are included in this PR? This PR upgrades ORC dependency from 1.9.2 to 2.0.0. ### Are these changes tested? Pass the CIs. ### Are there any user-facing changes? No. * GitHub Issue: #40507 Lead-authored-by: Antoine Pitrou <antoine@python.org> Co-authored-by: Gang Wu <ustcwg@gmail.com> Signed-off-by: Antoine Pitrou <antoine@python.org>
diff --git a/ci/scripts/python_wheel_macos_build.sh b/ci/scripts/python_wheel_macos_build.sh
@@ -91,6 +91,13 @@ echo "=== (${PYTHON_VERSION}) Building Arrow C++ libraries ==="
 : ${VCPKG_FEATURE_FLAGS:=-manifests}
 : ${VCPKG_TARGET_TRIPLET:=${VCPKG_DEFAULT_TRIPLET:-x64-osx-static-${CMAKE_BUILD_TYPE}}}
 
+echo "=== Protobuf compiler versions on PATH ==="
+which -a protoc || echo "no protoc on PATH!"
+
+echo "=== Protobuf compiler version from vcpkg ==="
+_pbc=${VCPKG_ROOT}/installed/${VCPKG_TARGET_TRIPLET}/tools/protobuf/protoc
+echo "$_pbc: `$_pbc --version`"
+
 mkdir -p ${build_dir}/build
 pushd ${build_dir}/build
 
@@ -122,6 +129,7 @@ cmake \
     -DARROW_SUBSTRAIT=${ARROW_SUBSTRAIT} \
     -DARROW_TENSORFLOW=${ARROW_TENSORFLOW} \
     -DARROW_USE_CCACHE=ON \
+    -DARROW_VERBOSE_THIRDPARTY_BUILD=ON \
     -DARROW_WITH_BROTLI=${ARROW_WITH_BROTLI} \
     -DARROW_WITH_BZ2=${ARROW_WITH_BZ2} \
     -DARROW_WITH_LZ4=${ARROW_WITH_LZ4} \
@@ -134,7 +142,6 @@ cmake \
     -DCMAKE_INSTALL_PREFIX=${build_dir}/install \
     -DCMAKE_OSX_ARCHITECTURES=${CMAKE_OSX_ARCHITECTURES} \
     -DCMAKE_UNITY_BUILD=${CMAKE_UNITY_BUILD} \
-    -DORC_PROTOBUF_EXECUTABLE=${VCPKG_ROOT}/installed/${VCPKG_TARGET_TRIPLET}/tools/protobuf/protoc \
     -DORC_SOURCE=BUNDLED \
     -DPARQUET_REQUIRE_ENCRYPTION=${PARQUET_REQUIRE_ENCRYPTION} \
     -DVCPKG_MANIFEST_MODE=OFF \
diff --git a/ci/scripts/python_wheel_manylinux_build.sh b/ci/scripts/python_wheel_manylinux_build.sh
@@ -123,7 +123,6 @@ cmake \
     -DCMAKE_INSTALL_LIBDIR=lib \
     -DCMAKE_INSTALL_PREFIX=/tmp/arrow-dist \
     -DCMAKE_UNITY_BUILD=${CMAKE_UNITY_BUILD} \
-    -DORC_PROTOBUF_EXECUTABLE=${VCPKG_ROOT}/installed/${VCPKG_TARGET_TRIPLET}/tools/protobuf/protoc \
     -DORC_SOURCE=BUNDLED \
     -DPARQUET_REQUIRE_ENCRYPTION=${PARQUET_REQUIRE_ENCRYPTION} \
     -DVCPKG_MANIFEST_MODE=OFF \
diff --git a/cpp/thirdparty/versions.txt b/cpp/thirdparty/versions.txt
@@ -90,8 +90,8 @@ ARROW_OPENTELEMETRY_BUILD_VERSION=v1.8.1
 ARROW_OPENTELEMETRY_BUILD_SHA256_CHECKSUM=3d640201594b07f08dade9cd1017bd0b59674daca26223b560b9bb6bf56264c2
 ARROW_OPENTELEMETRY_PROTO_BUILD_VERSION=v0.17.0
 ARROW_OPENTELEMETRY_PROTO_BUILD_SHA256_CHECKSUM=f269fbcb30e17b03caa1decd231ce826e59d7651c0f71c3b28eb5140b4bb5412
-ARROW_ORC_BUILD_VERSION=1.9.2
-ARROW_ORC_BUILD_SHA256_CHECKSUM=7f46f2c184ecefd6791f1a53fb062286818bd8710c3f08b94dd3cac365e240ee
+ARROW_ORC_BUILD_VERSION=2.0.0
+ARROW_ORC_BUILD_SHA256_CHECKSUM=9107730919c29eb39efaff1b9e36166634d1d4d9477e5fee76bfd6a8fec317df
 ARROW_PROTOBUF_BUILD_VERSION=v21.3
 ARROW_PROTOBUF_BUILD_SHA256_CHECKSUM=2f723218f6cb709ae4cdc4fb5ed56a5951fc5d466f0128ce4c946b8c78c8c49f
 # Because of https://github.com/Tencent/rapidjson/pull/1323, we require
diff --git a/dev/tasks/python-wheels/github.osx.yml b/dev/tasks/python-wheels/github.osx.yml
@@ -47,6 +47,10 @@ jobs:
           brew install bash bison coreutils ninja
           echo "$(brew --prefix bison)/bin" >> $GITHUB_PATH
 
+      - name: Homebrew packages
+        run: |
+          brew list
+
       - name: Retrieve VCPKG version from arrow/.env
         run: |
           vcpkg_version=$(cat "arrow/.env" | grep "VCPKG" | cut -d "=" -f2 | tr -d '"')
