[arrow] Update to 4.0.0 (#17975)

* Update arrow to 4.0.0

* Format

* Try fix thrift

* Update versions/ files

* Do not set ZSTD_ROOT

* Remove double quotes causing Windows problems

* Apply patches

* Remove LIB_DIR_OPTIONS

* Tweak zstd flags

* Update version hash

* Format

* Fail early on x86

* Update hash

* Fail early on arm, arm64

* Update hash

* Add expected failures to to scripts/ci.baseline.txt

* Exclude mallocs from default features

* Update hash

* Set default-features to false for aws-sdk-cpp

Co-authored-by: Robert Schumacher <roschuma@microsoft.com>

* Specify only x64 support in manifest

Co-authored-by: Robert Schumacher <roschuma@microsoft.com>

* Remove unneeded ci.baseline.txt entries

Co-authored-by: Robert Schumacher <roschuma@microsoft.com>

* Remove dataset from default-features

Co-authored-by: Robert Schumacher <roschuma@microsoft.com>

* Update hash

* Remove zstd path args

* Update hash

Co-authored-by: Tanguy Fautre <tanguy@fautre.com>
Co-authored-by: Robert Schumacher <roschuma@microsoft.com>

6 files changed, 168 insertions, 98 deletions

diff --git a/ports/arrow/CONTROL b/ports/arrow/CONTROL
deleted file mode 100644
index 8aace675f..000000000
--- a/ports/arrow/CONTROL
+++ /dev/null
@@ -1,20 +0,0 @@
-Source: arrow
-Version: 3.0.0
-Port-Version: 0
-Build-Depends: boost-algorithm, boost-filesystem, boost-multiprecision, boost-system, brotli, bzip2, double-conversion, flatbuffers, gflags, glog, lz4, openssl, rapidjson, re2, snappy, thrift, uriparser, utf8proc, zlib, zstd
-Homepage: https://github.com/apache/arrow
-Description: Apache Arrow is a columnar in-memory analytics layer designed to accelerate big data. It houses a set of canonical in-memory representations of flat and hierarchical data along with multiple language-bindings for structure manipulation. It also provides IPC and common algorithm implementations.
-Supports: x64&!x86&!arm
-Default-Features: csv, json, parquet, filesystem
-
-Feature: csv
-Description: CSV file support
-
-Feature: json
-Description: JSON file support
-
-Feature: parquet
-Description: Parquet file support
-
-Feature: filesystem
-Description: Local filesystem support
diff --git a/ports/arrow/all.patch b/ports/arrow/all.patch
index 838a9dcb4..8ff755ac5 100644
--- a/ports/arrow/all.patch
+++ b/ports/arrow/all.patch
@@ -1,11 +1,11 @@
 diff --git a/cpp/cmake_modules/BuildUtils.cmake b/cpp/cmake_modules/BuildUtils.cmake
-index e59b4a38a..9bd895608 100644
+index 2fd897b5d..b6118ad4f 100644
 --- a/cpp/cmake_modules/BuildUtils.cmake
 +++ b/cpp/cmake_modules/BuildUtils.cmake
 @@ -440,7 +440,7 @@ function(ADD_ARROW_LIB LIB_NAME)
        target_include_directories(${LIB_NAME}_static PRIVATE ${ARG_PRIVATE_INCLUDES})
      endif()
-
+ 
 -    if(MSVC_TOOLCHAIN)
 +    if(MSVC_TOOLCHAIN AND 0)
        set(LIB_NAME_STATIC ${LIB_NAME}_static)
@@ -26,26 +26,26 @@ index b46a0f1a0..3d87f5204 100644
 +  #pkg_check_modules(BROTLI_PC libbrotlicommon libbrotlienc libbrotlidec)
 +  if(BROTLI_PC_FOUND AND 0) # Find via pkg_check_modules disabled as incompatible with vcpkg
      set(BROTLI_INCLUDE_DIR "${BROTLI_PC_libbrotlicommon_INCLUDEDIR}")
-
+ 
      # Some systems (e.g. Fedora) don't fill Brotli_LIBRARY_DIRS, so add the other dirs here.
 diff --git a/cpp/cmake_modules/FindLz4.cmake b/cpp/cmake_modules/FindLz4.cmake
-index 14b6d93b9..d8d80c408 100644
+index 14b6d93b9..1905079ee 100644
 --- a/cpp/cmake_modules/FindLz4.cmake
 +++ b/cpp/cmake_modules/FindLz4.cmake
 @@ -15,10 +15,12 @@
  # specific language governing permissions and limitations
  # under the License.
-
+ 
 -if(MSVC_TOOLCHAIN AND NOT DEFINED LZ4_MSVC_LIB_PREFIX)
 -  set(LZ4_MSVC_LIB_PREFIX "lib")
 +# Avoid the debug build linking to the release library by mistake.
 +# In theory harmless if static linking at this point, but disastrous if done for a shared library.
 +if(CMAKE_BUILD_TYPE STREQUAL "DEBUG")
-+  set(LZ4_LIB_NAME_DEBUG_SUFFIX d)
++  set(LZ4_LIB_NAME_DEBUG_SUFFIX "d")
  endif()
 -set(LZ4_LIB_NAME_BASE "${LZ4_MSVC_LIB_PREFIX}lz4")
 +set(LZ4_LIB_NAME_BASE "lz4${LZ4_LIB_NAME_DEBUG_SUFFIX}")
-
+ 
  if(ARROW_LZ4_USE_SHARED)
    set(LZ4_LIB_NAMES)
 @@ -34,12 +36,8 @@ if(ARROW_LZ4_USE_SHARED)
@@ -60,10 +60,10 @@ index 14b6d93b9..d8d80c408 100644
 -      "${CMAKE_STATIC_LIBRARY_PREFIX}${LZ4_LIB_NAME_BASE}${LZ4_STATIC_LIB_SUFFIX}")
 +      "${CMAKE_STATIC_LIBRARY_PREFIX}${LZ4_LIB_NAME_BASE}${CMAKE_STATIC_LIBRARY_SUFFIX}")
  endif()
-
+ 
  if(LZ4_ROOT)
 @@ -56,8 +54,8 @@ if(LZ4_ROOT)
-
+ 
  else()
    find_package(PkgConfig QUIET)
 -  pkg_check_modules(LZ4_PC liblz4)
@@ -71,20 +71,20 @@ index 14b6d93b9..d8d80c408 100644
 +  #pkg_check_modules(LZ4_PC liblz4)
 +  if(0) # Do not use pkg_check_modules, doesn't seem to work correctly on some macOS versions (10.x in GitHub Actions)
      set(LZ4_INCLUDE_DIR "${LZ4_PC_INCLUDEDIR}")
-
+ 
      list(APPEND LZ4_PC_LIBRARY_DIRS "${LZ4_PC_LIBDIR}")
 diff --git a/cpp/cmake_modules/FindSnappy.cmake b/cpp/cmake_modules/FindSnappy.cmake
-index 5784cf592..817cf0c47 100644
+index 26cccb786..8bee097af 100644
 --- a/cpp/cmake_modules/FindSnappy.cmake
 +++ b/cpp/cmake_modules/FindSnappy.cmake
-@@ -15,20 +15,27 @@
+@@ -15,23 +15,30 @@
  # specific language governing permissions and limitations
  # under the License.
-
+ 
 +# Avoid the debug build linking to the release library by mistake.
 +# In theory harmless if static linking at this point, but disastrous if done for a shared library.
 +if(CMAKE_BUILD_TYPE STREQUAL "DEBUG")
-+  set(SNAPPY_LIB_NAME_DEBUG_SUFFIX d)
++  set(SNAPPY_LIB_NAME_DEBUG_SUFFIX "d")
 +endif()
 +
 +set(SNAPPY_LIB_NAME_BASE "snappy${SNAPPY_LIB_NAME_DEBUG_SUFFIX}")
@@ -102,26 +102,29 @@ index 5784cf592..817cf0c47 100644
  else()
 -  set(SNAPPY_STATIC_LIB_NAME_BASE "snappy")
    if(MSVC)
--    set(SNAPPY_STATIC_LIB_NAME_BASE "${SNAPPY_STATIC_LIB_NAME_BASE}${SNAPPY_MSVC_STATIC_LIB_SUFFIX}")
-+    set(SNAPPY_STATIC_LIB_NAME_BASE "${SNAPPY_LIB_NAME_BASE}${SNAPPY_MSVC_STATIC_LIB_SUFFIX}")
+     set(SNAPPY_STATIC_LIB_NAME_BASE
+-        "${SNAPPY_STATIC_LIB_NAME_BASE}${SNAPPY_MSVC_STATIC_LIB_SUFFIX}")
++        "${SNAPPY_LIB_NAME_BASE}${SNAPPY_MSVC_STATIC_LIB_SUFFIX}")
    endif()
--  set(SNAPPY_LIB_NAMES "${CMAKE_STATIC_LIBRARY_PREFIX}${SNAPPY_STATIC_LIB_NAME_BASE}${CMAKE_STATIC_LIBRARY_SUFFIX}")
-+  set(SNAPPY_LIB_NAMES "${CMAKE_STATIC_LIBRARY_PREFIX}${SNAPPY_LIB_NAME_BASE}${CMAKE_STATIC_LIBRARY_SUFFIX}")
+   set(
+     SNAPPY_LIB_NAMES
+-    "${CMAKE_STATIC_LIBRARY_PREFIX}${SNAPPY_STATIC_LIB_NAME_BASE}${CMAKE_STATIC_LIBRARY_SUFFIX}"
++    "${CMAKE_STATIC_LIBRARY_PREFIX}${SNAPPY_LIB_NAME_BASE}${CMAKE_STATIC_LIBRARY_SUFFIX}"
+     )
  endif()
-
- if(Snappy_ROOT)
+ 
 diff --git a/cpp/cmake_modules/FindThrift.cmake b/cpp/cmake_modules/FindThrift.cmake
-index 273d907ed..65f477f54 100644
+index 273d907ed..02a1e7fe1 100644
 --- a/cpp/cmake_modules/FindThrift.cmake
 +++ b/cpp/cmake_modules/FindThrift.cmake
 @@ -39,6 +39,12 @@ function(EXTRACT_THRIFT_VERSION)
    endif()
  endfunction(EXTRACT_THRIFT_VERSION)
-
+ 
 +# Avoid the debug build linking to the release library by mistake.
 +# In theory harmless if static linking at this point, but disastrous if done for a shared library.
 +if(CMAKE_BUILD_TYPE STREQUAL "DEBUG")
-+  set(THRIFT_LIB_NAME_DEBUG_SUFFIX d)
++  set(THRIFT_LIB_NAME_DEBUG_SUFFIX "d")
 +endif()
 +
  if(MSVC_TOOLCHAIN AND NOT DEFINED THRIFT_MSVC_LIB_SUFFIX)
@@ -133,7 +136,7 @@ index 273d907ed..65f477f54 100644
  endif()
 -set(THRIFT_LIB_NAME_BASE "thrift${THRIFT_MSVC_LIB_SUFFIX}")
 +set(THRIFT_LIB_NAME_BASE "thrift${THRIFT_MSVC_LIB_SUFFIX}${THRIFT_LIB_NAME_DEBUG_SUFFIX}")
-
+ 
  if(ARROW_THRIFT_USE_SHARED)
    set(THRIFT_LIB_NAMES thrift)
 @@ -84,8 +90,8 @@ else()
@@ -145,7 +148,7 @@ index 273d907ed..65f477f54 100644
 +  #pkg_check_modules(THRIFT_PC thrift)
 +  if(0) # Do not use pkg_check_modules, as it finds the wrong location (an intermediate build dir).
      set(THRIFT_INCLUDE_DIR "${THRIFT_PC_INCLUDEDIR}")
-
+ 
      list(APPEND THRIFT_PC_LIBRARY_DIRS "${THRIFT_PC_LIBDIR}")
 @@ -101,8 +107,7 @@ else()
      set(THRIFT_VERSION ${THRIFT_PC_VERSION})
@@ -157,26 +160,3 @@ index 273d907ed..65f477f54 100644
      find_path(THRIFT_INCLUDE_DIR thrift/Thrift.h PATH_SUFFIXES "include")
      find_program(THRIFT_COMPILER thrift PATH_SUFFIXES "bin")
      extract_thrift_version()
-diff --git a/cpp/cmake_modules/Findzstd.cmake b/cpp/cmake_modules/Findzstd.cmake
-index 6659a682d..d8cc4f72d 100644
---- a/cpp/cmake_modules/Findzstd.cmake
-+++ b/cpp/cmake_modules/Findzstd.cmake
-@@ -34,13 +34,14 @@ if(ARROW_ZSTD_USE_SHARED)
-       ZSTD_LIB_NAMES
-       "${CMAKE_SHARED_LIBRARY_PREFIX}${ZSTD_LIB_NAME_BASE}${CMAKE_SHARED_LIBRARY_SUFFIX}")
- else()
--  if(MSVC AND NOT DEFINED ZSTD_MSVC_STATIC_LIB_SUFFIX)
--    set(ZSTD_MSVC_STATIC_LIB_SUFFIX "_static")
-+  if(MSVC AND CMAKE_BUILD_TYPE STREQUAL "DEBUG")
-+    set(ZSTD_MSVC_DEBUG_LIB_SUFFIX d)
-   endif()
-   set(ZSTD_STATIC_LIB_SUFFIX
--      "${ZSTD_MSVC_STATIC_LIB_SUFFIX}${CMAKE_STATIC_LIBRARY_SUFFIX}")
-+      "${ZSTD_MSVC_DEBUG_LIB_SUFFIX}${CMAKE_STATIC_LIBRARY_SUFFIX}")
-   set(ZSTD_LIB_NAMES
--      "${CMAKE_STATIC_LIBRARY_PREFIX}${ZSTD_LIB_NAME_BASE}${ZSTD_STATIC_LIB_SUFFIX}")
-+      "zstd${ZSTD_STATIC_LIB_SUFFIX}"
-+      "libzstd${ZSTD_STATIC_LIB_SUFFIX}")
- endif()
-
- # First, find via if specified ZTD_ROOT
diff --git a/ports/arrow/portfile.cmake b/ports/arrow/portfile.cmake
index a36dad1c1..fb9e05bc6 100644
--- a/ports/arrow/portfile.cmake
+++ b/ports/arrow/portfile.cmake
@@ -3,53 +3,70 @@ vcpkg_fail_port_install(ON_ARCH "x86" "arm" "arm64")
 vcpkg_from_github(
     OUT_SOURCE_PATH SOURCE_PATH
     REPO apache/arrow
-    REF apache-arrow-3.0.0
-    SHA512 02645be0eaaaa69880ab911fc0b74665ebf52a35f9ad05210b23e7b42bcfbe3c3a4d44fa6c4c35af74764efbe528c2e0ebf0549ce5890c796be695ceb94e5606
+    REF apache-arrow-4.0.0
+    SHA512 4697a32004d02a519b8a8e899ed3cd981ae3485e6d34071436051080d6c84e25ad0bc568b3e52effe0a9204756da3d6e560a2037df06d2730dccd19c6b4c8027
     HEAD_REF master
     PATCHES
         all.patch
 )
 
-string(COMPARE EQUAL ${VCPKG_LIBRARY_LINKAGE} "dynamic" ARROW_BUILD_SHARED)
-string(COMPARE EQUAL ${VCPKG_LIBRARY_LINKAGE} "static" ARROW_BUILD_STATIC)
-
 vcpkg_check_features(OUT_FEATURE_OPTIONS FEATURE_OPTIONS
-    "csv"         ARROW_CSV
-    "json"        ARROW_JSON
-    "parquet"     ARROW_PARQUET
-    "filesystem"  ARROW_FILESYSTEM
+    FEATURES
+        csv         ARROW_CSV
+        dataset     ARROW_DATASET
+        filesystem  ARROW_FILESYSTEM
+        flight      ARROW_FLIGHT
+        json        ARROW_JSON
+        orc         ARROW_ORC
+        parquet     ARROW_PARQUET
+        parquet     PARQUET_REQUIRE_ENCRYPTION
+        s3          ARROW_S3
 )
 
-file(REMOVE "${SOURCE_PATH}/cpp/cmake_modules/FindZSTD.cmake")
+if(VCPKG_TARGET_IS_WINDOWS OR VCPKG_TARGET_IS_UWP)
+    set(MALLOC_OPTIONS -DARROW_JEMALLOC=OFF)
+elseif("jemalloc" IN_LIST FEATURES)
+    set(MALLOC_OPTIONS -DARROW_JEMALLOC=ON)
+else()
+    set(MALLOC_OPTIONS -DARROW_JEMALLOC=OFF)
+endif()
+
+if(VCPKG_TARGET_IS_WINDOWS AND ("mimalloc" IN_LIST FEATURES))
+    set(MALLOC_OPTIONS ${MALLOC_OPTIONS} -DARROW_MIMALLOC=ON)
+else()
+    set(MALLOC_OPTIONS ${MALLOC_OPTIONS} -DARROW_MIMALLOC=OFF)
+endif()
+
+string(COMPARE EQUAL ${VCPKG_LIBRARY_LINKAGE} "dynamic" ARROW_BUILD_SHARED)
+string(COMPARE EQUAL ${VCPKG_LIBRARY_LINKAGE} "static" ARROW_BUILD_STATIC)
+string(COMPARE EQUAL ${VCPKG_LIBRARY_LINKAGE} "dynamic" ARROW_DEPENDENCY_USE_SHARED)
+
+if(VCPKG_TARGET_IS_WINDOWS)
+    set(THRIFT_USE_SHARED OFF)
+else()
+    set(THRIFT_USE_SHARED ${ARROW_DEPENDENCY_USE_SHARED})
+endif()
 
 vcpkg_configure_cmake(
     SOURCE_PATH ${SOURCE_PATH}/cpp
     PREFER_NINJA
     OPTIONS
-        -DARROW_DEPENDENCY_SOURCE=SYSTEM
-        -Duriparser_SOURCE=SYSTEM
-        -DARROW_BUILD_TESTS=OFF
         ${FEATURE_OPTIONS}
-        -DARROW_BUILD_STATIC=${ARROW_BUILD_STATIC}
+        ${MALLOC_OPTIONS}
         -DARROW_BUILD_SHARED=${ARROW_BUILD_SHARED}
-        -DARROW_BROTLI_USE_SHARED=${ARROW_BUILD_SHARED}     # This can be wrong in custom triplets
-        -DARROW_GFLAGS_USE_SHARED=${ARROW_BUILD_SHARED}     # This can be wrong in custom triplets
-        -DARROW_LZ4_USE_SHARED=${ARROW_BUILD_SHARED}        # This can be wrong in custom triplets
-        -DARROW_SNAPPY_USE_SHARED=${ARROW_BUILD_SHARED}     # This can be wrong in custom triplets
-        -DARROW_THRIFT_USE_SHARED=OFF                       # vcpkg doesn't build Thrift as a shared library for the moment (2020/01/22).
-        -DARROW_UTF8PROC_USE_SHARED=${ARROW_BUILD_SHARED}   # This can be wrong in custom triplets
-        -DARROW_ZSTD_USE_SHARED=${ARROW_BUILD_SHARED}       # This can be wrong in custom triplets
-        -DARROW_JEMALLOC=OFF
-        -DARROW_BUILD_UTILITIES=OFF
+        -DARROW_BUILD_STATIC=${ARROW_BUILD_STATIC}
+        -DARROW_BUILD_TESTS=OFF
+        -DARROW_DEPENDENCY_SOURCE=SYSTEM
+        -DARROW_DEPENDENCY_USE_SHARED=${ARROW_DEPENDENCY_USE_SHARED}
+        -DARROW_THRIFT_USE_SHARED=${THRIFT_USE_SHARED} 
+        -DBUILD_WARNING_LEVEL=PRODUCTION
+        -DARROW_WITH_BROTLI=ON                 
         -DARROW_WITH_BZ2=ON
-        -DARROW_WITH_ZLIB=ON
-        -DARROW_WITH_ZSTD=ON
         -DARROW_WITH_LZ4=ON
         -DARROW_WITH_SNAPPY=ON
-        -DARROW_WITH_BROTLI=ON
-        -DARROW_WITH_UTF8PROC=ON
-        -DPARQUET_REQUIRE_ENCRYPTION=ON
-        -DBUILD_WARNING_LEVEL=PRODUCTION
+        -DARROW_WITH_ZLIB=ON
+        -DARROW_WITH_ZSTD=ON
+        -DZSTD_MSVC_LIB_PREFIX=
 )
 
 vcpkg_install_cmake()
diff --git a/ports/arrow/vcpkg.json b/ports/arrow/vcpkg.json
new file mode 100644
index 000000000..51755afd7
--- /dev/null
+++ b/ports/arrow/vcpkg.json
@@ -0,0 +1,88 @@
+{
+  "name": "arrow",
+  "version": "4.0.0",
+  "description": "Cross-language development platform for in-memory analytics",
+  "homepage": "https://arrow.apache.org",
+  "supports": "x64",
+  "dependencies": [
+    "boost-filesystem",
+    "boost-multiprecision",
+    "boost-system",
+    "brotli",
+    "bzip2",
+    "gflags",
+    "glog",
+    "lz4",
+    "openssl",
+    "re2",
+    "snappy",
+    "thrift",
+    "utf8proc",
+    "zlib",
+    "zstd"
+  ],
+  "default-features": [
+    "csv",
+    "filesystem",
+    "json",
+    "parquet"
+  ],
+  "features": {
+    "csv": {
+      "description": "CSV support"
+    },
+    "dataset": {
+      "description": "Dataset support"
+    },
+    "filesystem": {
+      "description": "Filesystem support"
+    },
+    "flight": {
+      "description": "Arrow Flight RPC support",
+      "dependencies": [
+        "abseil",
+        "c-ares",
+        "grpc",
+        "protobuf"
+      ]
+    },
+    "jemalloc": {
+      "description": "jemalloc allocator"
+    },
+    "json": {
+      "description": "JSON support",
+      "dependencies": [
+        "rapidjson"
+      ]
+    },
+    "mimalloc": {
+      "description": "mimalloc allocator"
+    },
+    "orc": {
+      "description": "ORC support",
+      "dependencies": [
+        "orc"
+      ]
+    },
+    "parquet": {
+      "description": "Parquet support"
+    },
+    "s3": {
+      "description": "S3 support",
+      "dependencies": [
+        {
+          "name": "aws-sdk-cpp",
+          "default-features": false,
+          "features": [
+            "cognito-identity",
+            "config",
+            "identity-management",
+            "s3",
+            "sts",
+            "transfer"
+          ]
+        }
+      ]
+    }
+  }
+}
diff --git a/versions/a-/arrow.json b/versions/a-/arrow.json
index d795d327a..503dbef98 100644
--- a/versions/a-/arrow.json
+++ b/versions/a-/arrow.json
@@ -1,6 +1,11 @@
 {
   "versions": [
     {
+      "git-tree": "2066704ab2b0b5977bcd5677ea1378fac0b0555c",
+      "version": "4.0.0",
+      "port-version": 0
+    },
+    {
       "git-tree": "bc9cb096ffa223f2af620db18ed244d17e47ffe7",
       "version-string": "3.0.0",
       "port-version": 0
diff --git a/versions/baseline.json b/versions/baseline.json
index 4ff7df52a..c46e2c457 100644
--- a/versions/baseline.json
+++ b/versions/baseline.json
@@ -157,7 +157,7 @@
       "port-version": 3
     },
     "arrow": {
-      "baseline": "3.0.0",
+      "baseline": "4.0.0",
       "port-version": 0
     },
     "ashes": {