WIP: Enable test suite of libarrow

recipe/bld.bat

@@ -20,12 +20,13 @@ cmake -G "Ninja" ^
       -DARROW_AZURE=%READ_RECIPE_META_YAML_WHY_NOT% ^
       -DARROW_BOOST_USE_SHARED:BOOL=ON ^
       -DARROW_BUILD_STATIC:BOOL=OFF ^
-      -DARROW_BUILD_TESTS:BOOL=OFF ^
+      -DARROW_BUILD_TESTS:BOOL=ON ^
       -DARROW_BUILD_UTILITIES:BOOL=OFF ^
       -DARROW_COMPUTE:BOOL=ON ^
       -DARROW_CSV:BOOL=ON ^
       -DARROW_DATASET:BOOL=ON ^
       -DARROW_DEPENDENCY_SOURCE=SYSTEM ^
+      -DARROW_ENABLE_TIMING_TESTS=OFF ^
       -DARROW_FILESYSTEM:BOOL=ON ^
       -DARROW_FLIGHT:BOOL=ON ^
       -DARROW_FLIGHT_REQUIRE_TLSCREDENTIALSOPTIONS:BOOL=ON ^
@@ -66,4 +67,10 @@ if %ERRORLEVEL% neq 0 exit 1
 cmake --build . --config Release
 if %ERRORLEVEL% neq 0 exit 1
 
+npm install -g azurite
+set ARROW_TEST_DATA=%SRC_DIR%\testing\data
+set PARQUET_TEST_DATA=%SRC_DIR%\cpp\submodules\parquet-testing\data
+ctest --progress --output-on-failure
+if %ERRORLEVEL% neq 0 exit 1
+
 popd

recipe/build.sh

@@ -57,13 +57,14 @@ cmake -GNinja \
     -DARROW_BOOST_USE_SHARED=ON \
     -DARROW_BUILD_BENCHMARKS=OFF \
     -DARROW_BUILD_STATIC=OFF \
-    -DARROW_BUILD_TESTS=OFF \
+    -DARROW_BUILD_TESTS=ON \
     -DARROW_BUILD_UTILITIES=OFF \
     -DARROW_COMPUTE=ON \
     -DARROW_CSV=ON \
     -DARROW_CXXFLAGS="${CXXFLAGS}" \
     -DARROW_DATASET=ON \
     -DARROW_DEPENDENCY_SOURCE=SYSTEM \
+    -DARROW_ENABLE_TIMING_TESTS=OFF \
     -DARROW_FILESYSTEM=ON \
     -DARROW_FLIGHT=ON \
     -DARROW_FLIGHT_REQUIRE_TLSCREDENTIALSOPTIONS=ON \
@@ -106,6 +107,13 @@ cmake -GNinja \
     ..
 
 # Do not install arrow, only build.
-cmake --build . --config Release
+cmake --build .
+
+if [[ "$CONDA_BUILD_CROSS_COMPILATION" != 1 ]]; then
+    npm install -g azurite
+    export ARROW_TEST_DATA=$SRC_DIR/testing/data
+    export PARQUET_TEST_DATA=$SRC_DIR/cpp/submodules/parquet-testing/data
+    ctest --progress --output-on-failure
+fi
 
 popd

recipe/meta.yaml

@@ -21,16 +21,27 @@ source:
       - patches/0001-GH-44455-C-Update-vendored-date-to-3.0.3.patch
       # backport https://github.com/apache/arrow/pull/44507
       - patches/0002-GH-44448-C-Add-support-for-overriding-grpc_cpp_plugi.patch
-  # testing-submodule not part of release tarball
+      # fix for https://github.com/apache/arrow/issues/43808
+      - patches/0003-GH-43808-C-set-kStrptimeSupportsZone-to-false-for-ol.patch
+      # skip gcsfs tests due to missing `storage-testbench`
+      - patches/0004-disable-gcsfs_test.patch
+  # testing-submodules not part of release tarball
   - git_url: https://github.com/apache/arrow-testing.git
     git_rev: 4d209492d514c2d3cb2d392681b9aa00e6d8da1c
     folder: testing
+  - git_url: https://github.com/apache/parquet-testing.git
+    git_rev: cb7a9674142c137367bf75a01b79c6e214a73199
+    folder: cpp/submodules/parquet-testing
 
 build:
   number: 0
   # for cuda support, building with one version is enough to be compatible with
   # all later versions, since arrow is only using libcuda, and not libcudart.
   skip: true  # [cuda_compiler_version not in ("None", cuda_compiler_version_min)]
+  # debug: only native compilation
+  skip: true  # [build_platform != target_platform]
+  # debug: cannot run CUDA-enabled libarrow tests without GPU
+  skip: true  # [cuda_compiler_version != "None"]
 
 requirements:
   build:
@@ -50,6 +61,8 @@ requirements:
     # necessary for vendored jemalloc
     - autoconf     # [unix]
     - make         # [unix]
+    # where orc ends up looking (because during build, CONDA_PREFIX==BUILD_PREFIX)
+    - tzdata
   host:
     # for required dependencies, see
     # https://github.com/apache/arrow/blob/apache-arrow-11.0.0/cpp/cmake_modules/ThirdpartyToolchain.cmake#L46-L75
@@ -73,7 +86,7 @@ requirements:
     # arrow uses a customized jemalloc, see #944
     # - jemalloc
     - libabseil
-    - libboost-headers
+    - libboost-devel *
     - libgoogle-cloud-devel
     - libgoogle-cloud-storage-devel
     - libgrpc
@@ -88,6 +101,15 @@ requirements:
     - xsimd
     - zlib
     - zstd
+    # test requirements, c.f.
+    # https://github.com/apache/arrow/blob/apache-arrow-12.0.0/cpp/CMakeLists.txt#L842-L846
+    - gtest
+    # https://github.com/apache/arrow/blob/apache-arrow-12.0.0/cpp/src/arrow/flight/sql/CMakeLists.txt#L86-L87
+    - sqlite *
+    # required by s3fs tests
+    - minio-server
+    # for npm, to install azurite
+    - nodejs *
 
 outputs:
   - name: apache-arrow-proc

recipe/patches/0001-GH-44455-C-Update-vendored-date-to-3.0.3.patch

@@ -1,7 +1,7 @@
 From d774a80949599d58d4726aaa2d78cc64208ab300 Mon Sep 17 00:00:00 2001
 From: Sutou Kouhei <[email protected]>
 Date: Mon, 21 Oct 2024 16:41:27 +0900
-Subject: [PATCH 1/2] GH-44455: [C++] Update vendored date to 3.0.3
+Subject: [PATCH 1/4] GH-44455: [C++] Update vendored date to 3.0.3
 
 IANA tzdata changed its data format. So we need to update vendored
 date to parse it.

recipe/patches/0002-GH-44448-C-Add-support-for-overriding-grpc_cpp_plugi.patch

@@ -1,7 +1,7 @@
 From 3d00b553ffd302ada8d74b5b77b71128c594c0b4 Mon Sep 17 00:00:00 2001
 From: Sutou Kouhei <[email protected]>
 Date: Wed, 23 Oct 2024 10:30:33 +0900
-Subject: [PATCH 2/2] GH-44448: [C++] Add support for overriding
+Subject: [PATCH 2/4] GH-44448: [C++] Add support for overriding
  grpc_cpp_plugin path for cross-compiling
 
 ---

recipe/patches/0003-GH-43808-C-set-kStrptimeSupportsZone-to-false-for-ol.patch

@@ -0,0 +1,24 @@
+From 9c482984116a6e6dcce7e9f91af453aa51b966ee Mon Sep 17 00:00:00 2001
+From: "H. Vetinari" <[email protected]>
+Date: Sat, 2 Nov 2024 15:29:20 +1100
+Subject: [PATCH 3/4] GH-43808: [C++] set kStrptimeSupportsZone to false for
+ old glibc
+
+---
+ cpp/src/arrow/util/value_parsing.h | 3 +++
+ 1 file changed, 3 insertions(+)
+
+diff --git a/cpp/src/arrow/util/value_parsing.h b/cpp/src/arrow/util/value_parsing.h
+index 609906052c..d6a615f6f6 100644
+--- a/cpp/src/arrow/util/value_parsing.h
++++ b/cpp/src/arrow/util/value_parsing.h
+@@ -790,6 +790,9 @@ static inline bool ParseTimestampISO8601(const char* s, size_t length,
+
+ #if defined(_WIN32) || defined(ARROW_WITH_MUSL)
+ static constexpr bool kStrptimeSupportsZone = false;
++#elif defined(__GLIBC__) && defined(__GLIBC_MINOR__) && (__GLIBC__ == 2) && (__GLIBC_MINOR__ < 28)
++// see https://github.com/apache/arrow/issues/43808
++static constexpr bool kStrptimeSupportsZone = false;
+ #else
+ static constexpr bool kStrptimeSupportsZone = true;
+ #endif

recipe/patches/0004-disable-gcsfs_test.patch

@@ -0,0 +1,31 @@
+From 43e073eb4821d1e940915add71553c84028f3358 Mon Sep 17 00:00:00 2001
+From: "H. Vetinari" <[email protected]>
+Date: Sat, 2 Nov 2024 15:41:34 +1100
+Subject: [PATCH 4/4] disable gcsfs_test
+
+it cannot work unless we package https://github.com/googleapis/storage-testbench,
+which however has extremely tight dependencies on protobuf etc., making it very
+hard to fit this into our migration patterns
+---
+ cpp/src/arrow/filesystem/CMakeLists.txt | 8 --------
+ 1 file changed, 8 deletions(-)
+
+diff --git a/cpp/src/arrow/filesystem/CMakeLists.txt b/cpp/src/arrow/filesystem/CMakeLists.txt
+index 7afdf566f2..b22a48a0c9 100644
+--- a/cpp/src/arrow/filesystem/CMakeLists.txt
++++ b/cpp/src/arrow/filesystem/CMakeLists.txt
+@@ -42,14 +42,6 @@ if(ARROW_BUILD_BENCHMARKS)
+                       ${ARROW_BENCHMARK_LINK_LIBS})
+ endif()
+
+-if(ARROW_GCS)
+-  add_arrow_test(gcsfs_test
+-                 EXTRA_LABELS
+-                 filesystem
+-                 EXTRA_LINK_LIBS
+-                 google-cloud-cpp::storage)
+-endif()
+-
+ if(ARROW_AZURE)
+   add_arrow_test(azurefs_test
+                  EXTRA_LABELS