Merge branch 'apache:main' into Pyarrow_WoA

Author: MugundanMCW

.env

@@ -84,7 +84,6 @@ R_ORG=rhub
 R_TAG=latest
 
 # Env vars for R builds
-R_UPDATE_CLANG=false
 R_CUSTOM_CCACHE=false
 ARROW_R_DEV=TRUE
 R_PRUNE_DEPS=FALSE

.github/workflows/cpp_extra.yml

@@ -107,6 +107,16 @@ jobs:
           - image: alpine-linux-cpp
             runs-on: ubuntu-latest
             title: AMD64 Alpine Linux
+          - image: ubuntu-cpp
+            run-options: >-
+              -e ARROW_CTEST_TIMEOUT=2000
+              -e ARROW_C_FLAGS_DEBUG="-O1"
+              -e ARROW_CXX_FLAGS_DEBUG="-O1"
+              -e ARROW_GANDIVA=OFF
+              -e ARROW_LARGE_MEMORY_TESTS=ON
+              -e BUILD_WARNING_LEVEL=PRODUCTION
+            runs-on: "runs-on=${{ github.run_id }}/family=x8i.2xlarge/volume=80gb/spot=capacity-optimized"
+            title: AMD64 Ubuntu Large Memory Tests
           - image: conda-cpp
             run-options: >-
               -e ARROW_USE_MESON=ON
@@ -336,6 +346,62 @@ jobs:
           cd cpp/examples/minimal_build
           ../minimal_build.build/arrow-example
 
+  odbc-linux:
+    needs: check-labels
+    name: ODBC Linux
+    runs-on: ubuntu-latest
+    if: >-
+      needs.check-labels.outputs.force == 'true' ||
+      contains(fromJSON(needs.check-labels.outputs.ci-extra-labels || '[]'), 'CI: Extra') ||
+      contains(fromJSON(needs.check-labels.outputs.ci-extra-labels || '[]'), 'CI: Extra: C++')
+    timeout-minutes: 75
+    strategy:
+      fail-fast: false
+    env:
+      ARCH: amd64
+      ARCHERY_DEBUG: 1
+      ARROW_ENABLE_TIMING_TESTS: OFF
+      DOCKER_VOLUME_PREFIX: ".docker/"
+      UBUNTU: 24.04
+    steps:
+      - name: Checkout Arrow
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+          submodules: recursive
+      - name: Cache Docker Volumes
+        uses: actions/cache@v5
+        with:
+          path: .docker
+          key: ubuntu-cpp-odbc-${{ hashFiles('cpp/**') }}
+          restore-keys: ubuntu-cpp-odbc-
+      - name: Setup Python on hosted runner
+        uses: actions/setup-python@v6
+        with:
+          python-version: 3
+      - name: Setup Archery
+        run: python3 -m pip install -e dev/archery[docker]
+      - name: Execute Docker Build
+        env:
+          ARCHERY_DOCKER_USER: ${{ secrets.DOCKERHUB_USER }}
+          ARCHERY_DOCKER_PASSWORD: ${{ secrets.DOCKERHUB_TOKEN }}
+        run: |
+          # GH-40558: reduce ASLR to avoid ASAN/LSAN crashes
+          sudo sysctl -w vm.mmap_rnd_bits=28
+          source ci/scripts/util_enable_core_dumps.sh
+          archery docker run ubuntu-cpp-odbc
+      - name: Docker Push
+        if: >-
+          success() &&
+          github.event_name == 'push' &&
+          github.repository == 'apache/arrow' &&
+          github.ref_name == 'main'
+        env:
+          ARCHERY_DOCKER_USER: ${{ secrets.DOCKERHUB_USER }}
+          ARCHERY_DOCKER_PASSWORD: ${{ secrets.DOCKERHUB_TOKEN }}
+        continue-on-error: true
+        run: archery docker push ubuntu-cpp-odbc
+
   odbc-macos:
     needs: check-labels
     name: ODBC ${{ matrix.build-type }} ${{ matrix.architecture }} macOS ${{ matrix.macos-version }}
@@ -435,7 +501,7 @@ jobs:
             "$(pwd)/build/cpp/${{ matrix.build-type }}/libarrow_flight_sql_odbc.dylib"
       - name: Register Flight SQL ODBC Driver
         run: |
-          sudo cpp/src/arrow/flight/sql/odbc/install/mac/install_odbc.sh $(pwd)/build/cpp/${{ matrix.build-type }}/libarrow_flight_sql_odbc.dylib
+          sudo cpp/src/arrow/flight/sql/odbc/install/unix/install_odbc.sh $(pwd)/build/cpp/${{ matrix.build-type }}/libarrow_flight_sql_odbc.dylib
       - name: Test
         shell: bash
         run: |
@@ -698,6 +764,7 @@ jobs:
       - jni-linux
       - jni-macos
       - msvc-arm64
+      - odbc-linux
       - odbc-macos
       - odbc-msvc
       - odbc-nightly

.pre-commit-config.yaml

@@ -353,7 +353,7 @@ repos:
           ?^cpp/build-support/update-thrift\.sh$|
           ?^cpp/examples/minimal_build/run\.sh$|
           ?^cpp/examples/tutorial_examples/run\.sh$|
-          ?^cpp/src/arrow/flight/sql/odbc/install/mac/install_odbc\.sh$|
+          ?^cpp/src/arrow/flight/sql/odbc/install/unix/install_odbc\.sh$|
           ?^dev/release/05-binary-upload\.sh$|
           ?^dev/release/08-binary-verify\.sh$|
           ?^dev/release/binary-recover\.sh$|

ci/docker/linux-r.dockerfile

@@ -33,9 +33,6 @@ ENV R_PRUNE_DEPS=${r_prune_deps}
 ARG r_custom_ccache=false
 ENV R_CUSTOM_CCACHE=${r_custom_ccache}
 
-ARG r_update_clang=false
-ENV R_UPDATE_CLANG=${r_update_clang}
-
 ARG tz="UTC"
 ENV TZ=${tz}
 

ci/docker/ubuntu-24.04-cpp.dockerfile

@@ -121,6 +121,7 @@ RUN apt-get update -y -q && \
         rsync \
         tzdata \
         tzdata-legacy \
+        unixodbc-dev \
         uuid-runtime \
         unzip \
         wget && \

ci/scripts/r_docker_configure.sh

@@ -85,17 +85,6 @@ else
   $PACKAGE_MANAGER install -y rsync cmake curl
 fi
 
-# Update clang version to latest available.
-# This is only for rhub/clang20. If we change the base image from rhub/clang20,
-# we need to update this part too.
-if [ "$R_UPDATE_CLANG" = true ]; then
-  apt update -y --allow-releaseinfo-change # flag needed for when debian version changes
-  apt install -y gnupg
-  curl -fsSL https://apt.llvm.org/llvm-snapshot.gpg.key | gpg --dearmor -o /etc/apt/trusted.gpg.d/llvm.gpg
-  echo "deb http://apt.llvm.org/jammy/ llvm-toolchain-jammy-20 main" > /etc/apt/sources.list.d/llvm20.list
-  apt update -y --allow-releaseinfo-change # flag needed for when debian version changes
-  apt install -y clang-20 lld-20
-fi
 
 # Workaround for html help install failure; see https://github.com/r-lib/devtools/issues/2084#issuecomment-530912786
 Rscript -e 'x <- file.path(R.home("doc"), "html"); if (!file.exists(x)) {dir.create(x, recursive=TRUE); file.copy(system.file("html/R.css", package="stats"), x)}'

ci/scripts/r_install_system_dependencies.sh

@@ -35,17 +35,19 @@ else
   apt-get update
 fi
 
-# Install curl and OpenSSL (technically, only needed for S3/GCS support, but
-# installing the R curl package fails without it)
+# Install curl, OpenSSL, and libuv
+# - curl/OpenSSL: technically only needed for S3/GCS support, but
+#   installing the R curl package fails without it
+# - libuv: required by the fs R package (no longer bundles libuv by default)
 case "$PACKAGE_MANAGER" in
   apt-get)
-    apt-get install -y libcurl4-openssl-dev libssl-dev
+    apt-get install -y libcurl4-openssl-dev libssl-dev libuv1-dev
     ;;
   apk)
-    $PACKAGE_MANAGER add curl-dev openssl-dev
+    $PACKAGE_MANAGER add curl-dev openssl-dev libuv-dev
     ;;
   *)
-    $PACKAGE_MANAGER install -y libcurl-devel openssl-devel
+    $PACKAGE_MANAGER install -y libcurl-devel openssl-devel libuv-devel
     ;;
 esac
 

ci/scripts/r_revdepcheck.sh

@@ -36,6 +36,7 @@ apt install -y \
   libbz2-dev \
   libc-ares-dev \
   libcurl4-openssl-dev \
+  libuv1-dev \
   libgflags-dev \
   libgoogle-glog-dev \
   liblz4-dev \

compose.yaml

@@ -151,6 +151,7 @@ x-hierarchy:
     - ubuntu-r-only-r
   - ubuntu-cpp-bundled
   - ubuntu-cpp-bundled-offline
+  - ubuntu-cpp-odbc
   - ubuntu-cpp-minimal
   - ubuntu-cuda-cpp:
     - ubuntu-cuda-python
@@ -371,7 +372,7 @@ services:
         /arrow/ci/scripts/cpp_build.sh /arrow /build &&
         /arrow/ci/scripts/cpp_test.sh /arrow /build"
 
-  ubuntu-cpp:
+  ubuntu-cpp: &ubuntu-cpp-base
     # Usage:
     #   docker compose build ubuntu-cpp
     #   docker compose run --rm ubuntu-cpp
@@ -496,6 +497,33 @@ services:
     volumes: *ubuntu-volumes
     command: *cpp-command
 
+  ubuntu-cpp-odbc:
+    # Arrow Flight SQL ODBC build with BUNDLED dependencies with downloaded dependencies.
+    <<: *ubuntu-cpp-base
+    environment:
+      <<: [*common, *ccache, *sccache, *cpp]
+      ARROW_ACERO: "OFF"
+      ARROW_AZURE: "OFF"
+      ARROW_BUILD_TYPE: RELEASE
+      ARROW_CSV: "OFF"
+      ARROW_DATASET: "OFF"
+      ARROW_DEPENDENCY_SOURCE: BUNDLED
+      ARROW_DEPENDENCY_USE_SHARED: "OFF"
+      ARROW_FLIGHT_SQL_ODBC: "ON"
+      ARROW_GANDIVA: "OFF"
+      ARROW_GCS: "OFF"
+      ARROW_HDFS: "OFF"
+      ARROW_ORC: "OFF"
+      ARROW_PARQUET: "OFF"
+      ARROW_S3: "OFF"
+      ARROW_SUBSTRAIT: "OFF"
+    # Register ODBC before running tests
+    command: >
+      /bin/bash -c "
+        /arrow/ci/scripts/cpp_build.sh /arrow /build &&
+        sudo /arrow/cpp/src/arrow/flight/sql/odbc/install/unix/install_odbc.sh /usr/local/lib/libarrow_flight_sql_odbc.so &&
+        /arrow/ci/scripts/cpp_test.sh /arrow /build"
+
   ubuntu-cpp-minimal:
     # Arrow build with minimal components/dependencies
     image: ${REPO}:${ARCH}-ubuntu-${UBUNTU}-cpp-minimal
@@ -1667,7 +1695,6 @@ services:
         tz: ${TZ}
         r_prune_deps: ${R_PRUNE_DEPS}
         r_custom_ccache: ${R_CUSTOM_CCACHE}
-        r_update_clang: ${R_UPDATE_CLANG}
     shm_size: *shm-size
     environment:
       <<: [*common, *sccache]

cpp/CMakeLists.txt

@@ -211,17 +211,6 @@ else()
   set(MSVC_TOOLCHAIN FALSE)
 endif()
 
-find_package(ClangTools)
-find_package(InferTools)
-if("$ENV{CMAKE_EXPORT_COMPILE_COMMANDS}" STREQUAL "1"
-   OR CLANG_TIDY_FOUND
-   OR INFER_FOUND)
-  # Generate a Clang compile_commands.json "compilation database" file for use
-  # with various development tools, such as Vim's YouCompleteMe plugin.
-  # See http://clang.llvm.org/docs/JSONCompilationDatabase.html
-  set(CMAKE_EXPORT_COMPILE_COMMANDS 1)
-endif()
-
 # Needed for Gandiva.
 # Use the first Python installation on PATH, not the newest one
 set(Python3_FIND_STRATEGY "LOCATION")
@@ -649,22 +638,6 @@ if(UNIX)
                     VERBATIM)
 endif(UNIX)
 
-#
-# "make infer" target
-#
-
-if(${INFER_FOUND})
-  # runs infer capture
-  add_custom_target(infer ${BUILD_SUPPORT_DIR}/run-infer.sh ${INFER_BIN}
-                          ${CMAKE_BINARY_DIR}/compile_commands.json 1)
-  # runs infer analyze
-  add_custom_target(infer-analyze ${BUILD_SUPPORT_DIR}/run-infer.sh ${INFER_BIN}
-                                  ${CMAKE_BINARY_DIR}/compile_commands.json 2)
-  # runs infer report
-  add_custom_target(infer-report ${BUILD_SUPPORT_DIR}/run-infer.sh ${INFER_BIN}
-                                 ${CMAKE_BINARY_DIR}/compile_commands.json 3)
-endif()
-
 #
 # Link targets
 #