early-access version 2853

externals/vcpkg/ports/arrow/all.patch

@@ -0,0 +1,13 @@
+diff --git a/cpp/cmake_modules/BuildUtils.cmake b/cpp/cmake_modules/BuildUtils.cmake
+index 391c43e0a..50f6d3d3c 100644
+--- a/cpp/cmake_modules/BuildUtils.cmake
++++ b/cpp/cmake_modules/BuildUtils.cmake
+@@ -427,7 +427,7 @@ 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)
+     else()
+       set(LIB_NAME_STATIC ${LIB_NAME})

externals/vcpkg/ports/arrow/fix-ThirdPartyToolchain.patch

@@ -0,0 +1,143 @@
+diff --git a/cpp/cmake_modules/ThirdpartyToolchain.cmake b/cpp/cmake_modules/ThirdpartyToolchain.cmake
+index f070323..290b5a7 100644
+--- a/cpp/cmake_modules/ThirdpartyToolchain.cmake
++++ b/cpp/cmake_modules/ThirdpartyToolchain.cmake
+@@ -53,7 +53,7 @@ set(ARROW_THIRDPARTY_DEPENDENCIES
+     AWSSDK
+     benchmark
+     Boost
+-    Brotli
++    unofficial-brotli
+     BZip2
+     c-ares
+     gflags
+@@ -62,7 +62,7 @@ set(ARROW_THIRDPARTY_DEPENDENCIES
+     gRPC
+     GTest
+     LLVM
+-    Lz4
++    lz4
+     nlohmann_json
+     opentelemetry-cpp
+     ORC
+@@ -72,7 +72,7 @@ set(ARROW_THIRDPARTY_DEPENDENCIES
+     Snappy
+     Substrait
+     Thrift
+-    utf8proc
++    unofficial-utf8proc
+     xsimd
+     ZLIB
+     zstd)
+@@ -139,6 +139,9 @@ foreach(DEPENDENCY ${ARROW_THIRDPARTY_DEPENDENCIES})
+ endforeach()
+ 
+ macro(build_dependency DEPENDENCY_NAME)
++  if(NOT "${DEPENDENCY_NAME}" STREQUAL "xsimd")
++    message(FATAL_ERROR "dependencies should never be built in vcpkg")
++  endif()
+   if("${DEPENDENCY_NAME}" STREQUAL "AWSSDK")
+     build_awssdk()
+   elseif("${DEPENDENCY_NAME}" STREQUAL "benchmark")
+@@ -222,7 +225,7 @@ macro(resolve_dependency DEPENDENCY_NAME)
+   endif()
+ 
+   if(ARG_HAVE_ALT)
+-    set(PACKAGE_NAME "${DEPENDENCY_NAME}Alt")
++    set(PACKAGE_NAME "${DEPENDENCY_NAME}")
+   else()
+     set(PACKAGE_NAME ${DEPENDENCY_NAME})
+   endif()
+@@ -1129,10 +1132,16 @@ macro(build_brotli)
+ endmacro()
+ 
+ if(ARROW_WITH_BROTLI)
+-  resolve_dependency(Brotli PC_PACKAGE_NAMES libbrotlidec libbrotlienc)
++  resolve_dependency(unofficial-brotli USE_CONFIG TRUE PC_PACKAGE_NAMES libbrotlidec libbrotlienc)
++  add_library(Brotli::brotlicommon ALIAS unofficial::brotli::brotlicommon)
++  add_library(Brotli::brotlienc ALIAS unofficial::brotli::brotlienc)
++  add_library(Brotli::brotlidec ALIAS unofficial::brotli::brotlidec)
++  get_target_property(BROTLI_INCLUDE_DIR unofficial::brotli::brotlicommon-static INTERFACE_INCLUDE_DIRECTORIES)
+   # TODO: Don't use global includes but rather target_include_directories
++  if(NOT BROTLI_INCLUDE_DIR)
+   get_target_property(BROTLI_INCLUDE_DIR Brotli::brotlicommon
+                       INTERFACE_INCLUDE_DIRECTORIES)
++  endif()
+   include_directories(SYSTEM ${BROTLI_INCLUDE_DIR})
+ endif()
+ 
+@@ -1249,7 +1261,7 @@ macro(build_glog)
+ endmacro()
+ 
+ if(ARROW_USE_GLOG)
+-  resolve_dependency(GLOG PC_PACKAGE_NAMES libglog)
++  resolve_dependency(glog USE_CONFIG TRUE PC_PACKAGE_NAMES libglog)
+   # TODO: Don't use global includes but rather target_include_directories
+   get_target_property(GLOG_INCLUDE_DIR glog::glog INTERFACE_INCLUDE_DIRECTORIES)
+   include_directories(SYSTEM ${GLOG_INCLUDE_DIR})
+@@ -1434,6 +1446,13 @@ if(ARROW_WITH_THRIFT)
+   # TODO: Don't use global includes but rather target_include_directories
+   include_directories(SYSTEM ${THRIFT_INCLUDE_DIR})
+ 
++  if(THRIFT_INCLUDE_DIR)
++    file(READ "${THRIFT_INCLUDE_DIR}/config.h" THRIFT_CONFIG_H_CONTENT)
++    string(REGEX MATCH "#define PACKAGE_VERSION \"[0-9.]+\"" THRIFT_VERSION_DEFINITION "${THRIFT_CONFIG_H_CONTENT}")
++    string(REGEX MATCH "[0-9.]+" THRIFT_VERSION "${THRIFT_VERSION_DEFINITION}")
++    set(THRIFT_VERSION "${THRIFT_VERSION}")
++  endif()
++
+   string(REPLACE "." ";" VERSION_LIST ${THRIFT_VERSION})
+   list(GET VERSION_LIST 0 THRIFT_VERSION_MAJOR)
+   list(GET VERSION_LIST 1 THRIFT_VERSION_MINOR)
+@@ -1561,6 +1580,8 @@ if(ARROW_WITH_PROTOBUF)
+                      ${ARROW_PROTOBUF_REQUIRED_VERSION}
+                      PC_PACKAGE_NAMES
+                      protobuf)
++  get_target_property(PROTOBUF_INCLUDE_DIR protobuf::libprotobuf
++                      INTERFACE_INCLUDE_DIRECTORIES)
+ 
+   if(NOT Protobuf_USE_STATIC_LIBS AND MSVC_TOOLCHAIN)
+     add_definitions(-DPROTOBUF_USE_DLLS)
+@@ -2121,8 +2142,6 @@ endmacro()
+ if(ARROW_WITH_RAPIDJSON)
+   set(ARROW_RAPIDJSON_REQUIRED_VERSION "1.1.0")
+   resolve_dependency(RapidJSON
+-                     HAVE_ALT
+-                     TRUE
+                      REQUIRED_VERSION
+                      ${ARROW_RAPIDJSON_REQUIRED_VERSION}
+                      IS_RUNTIME_DEPENDENCY
+@@ -2260,7 +2279,8 @@ macro(build_lz4)
+ endmacro()
+ 
+ if(ARROW_WITH_LZ4)
+-  resolve_dependency(Lz4 PC_PACKAGE_NAMES liblz4)
++  resolve_dependency(lz4 USE_CONFIG TRUE PC_PACKAGE_NAMES liblz4)
++  add_library(LZ4::lz4 ALIAS lz4::lz4)
+ 
+   # TODO: Don't use global includes but rather target_include_directories
+   get_target_property(LZ4_INCLUDE_DIR LZ4::lz4 INTERFACE_INCLUDE_DIRECTORIES)
+@@ -2517,11 +2537,10 @@ macro(build_utf8proc)
+ endmacro()
+ 
+ if(ARROW_WITH_UTF8PROC)
+-  resolve_dependency(utf8proc
+-                     REQUIRED_VERSION
+-                     "2.2.0"
++  resolve_dependency(unofficial-utf8proc
+                      PC_PACKAGE_NAMES
+                      libutf8proc)
++  add_library(utf8proc::utf8proc ALIAS utf8proc)
+ 
+   add_definitions(-DARROW_WITH_UTF8PROC)
+ 
+@@ -3719,6 +3738,9 @@ if(ARROW_WITH_GRPC)
+   # TODO: Don't use global includes but rather target_include_directories
+   get_target_property(GRPC_INCLUDE_DIR gRPC::grpc++ INTERFACE_INCLUDE_DIRECTORIES)
+   include_directories(SYSTEM ${GRPC_INCLUDE_DIR})
++  if (ABSL_USE_CXX17)
++    set(CMAKE_CXX_STANDARD 17)
++  endif()
+ 
+   if(GRPC_VENDORED)
+     set(GRPCPP_PP_INCLUDE TRUE)

externals/vcpkg/ports/arrow/portfile.cmake

@@ -0,0 +1,118 @@
+vcpkg_from_github(
+    OUT_SOURCE_PATH SOURCE_PATH
+    REPO apache/arrow
+    REF apache-arrow-8.0.0
+    SHA512 08b6ab4a3c5e0dd9c46402da8e7b9ef9f918eea177413cb31695192dfdb5a472ebbfef255b8343fe775d81e8b5eb268c3428a699fac48b36bf808f5b81e83a64
+    HEAD_REF master
+    PATCHES
+        vs-2022-fixes.patch
+        all.patch
+        fix-ThirdPartyToolchain.patch
+)
+file(REMOVE "${SOURCE_PATH}/cpp/cmake_modules/Findzstd.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/FindBrotli.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/Find-c-aresAlt.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/FindLz4.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/FindSnappy.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/FindThrift.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/FindGLOG.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/Findutf8proc.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/FindRapidJSONAlt.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/FindgRPCAlt.cmake"
+            "${SOURCE_PATH}/cpp/cmake_modules/FindgflagsAlt.cmake"
+)
+
+vcpkg_check_features(OUT_FEATURE_OPTIONS FEATURE_OPTIONS
+    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
+        plasma      ARROW_PLASMA
+        s3          ARROW_S3
+        cuda        ARROW_CUDA
+)
+
+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 ${ARROW_DEPENDENCY_USE_SHARED})
+endif()
+
+vcpkg_cmake_configure(
+    SOURCE_PATH "${SOURCE_PATH}/cpp"
+    OPTIONS
+        ${FEATURE_OPTIONS}
+        ${MALLOC_OPTIONS}
+        ${S3_OPTIONS}
+        -DCMAKE_SYSTEM_PROCESSOR=${VCPKG_TARGET_ARCHITECTURE}
+        -DARROW_BUILD_SHARED=${ARROW_BUILD_SHARED}
+        -DARROW_BUILD_STATIC=${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_LZ4=ON
+        -DARROW_WITH_SNAPPY=ON
+        -DARROW_WITH_ZLIB=ON
+        -DARROW_WITH_ZSTD=ON
+        -DZSTD_MSVC_LIB_PREFIX=
+    MAYBE_UNUSED_VARIABLES
+        ZSTD_MSVC_LIB_PREFIX
+)
+
+vcpkg_cmake_install()
+
+vcpkg_copy_pdbs()
+
+if(EXISTS "${CURRENT_PACKAGES_DIR}/lib/arrow_static.lib")
+    message(FATAL_ERROR "Installed lib file should be named 'arrow.lib' via patching the upstream build.")
+endif()
+
+vcpkg_cmake_config_fixup(CONFIG_PATH lib/cmake/arrow)
+
+file(REMOVE_RECURSE "${CURRENT_PACKAGES_DIR}/debug/lib/cmake")
+file(REMOVE_RECURSE "${CURRENT_PACKAGES_DIR}/lib/cmake")
+
+configure_file(${CMAKE_CURRENT_LIST_DIR}/vcpkg-cmake-wrapper.cmake ${CURRENT_PACKAGES_DIR}/share/${PORT} @ONLY)
+file(INSTALL "${CMAKE_CURRENT_LIST_DIR}/usage" DESTINATION "${CURRENT_PACKAGES_DIR}/share/${PORT}")
+file(INSTALL "${SOURCE_PATH}/LICENSE.txt" DESTINATION "${CURRENT_PACKAGES_DIR}/share/${PORT}" RENAME copyright)
+
+file(GLOB PARQUET_FILES ${CURRENT_PACKAGES_DIR}/share/${PORT}/Parquet*)
+file(COPY ${PARQUET_FILES} DESTINATION "${CURRENT_PACKAGES_DIR}/share/parquet")
+file(REMOVE_RECURSE ${PARQUET_FILES})
+file(RENAME ${CURRENT_PACKAGES_DIR}/share/${PORT}/FindParquet.cmake ${CURRENT_PACKAGES_DIR}/share/parquet/FindParquet.cmake)
+
+file(REMOVE_RECURSE "${CURRENT_PACKAGES_DIR}/debug/include")
+file(REMOVE_RECURSE "${CURRENT_PACKAGES_DIR}/debug/share")
+
+if ("plasma" IN_LIST FEATURES)
+    vcpkg_copy_tools(TOOL_NAMES plasma-store-server AUTO_CLEAN)
+endif ()
+
+vcpkg_fixup_pkgconfig()

externals/vcpkg/ports/arrow/usage

@@ -0,0 +1,7 @@
+The package arrow provides CMake targets:
+
+    find_package(Arrow CONFIG REQUIRED)
+    target_link_libraries(main PRIVATE ${ARROW_LIBRARIES})
+
+    find_package(Parquet CONFIG REQUIRED)
+    target_link_libraries(main PRIVATE ${PARQUET_LIBRARIES})

externals/vcpkg/ports/arrow/vcpkg-cmake-wrapper.cmake

@@ -0,0 +1,9 @@
+_find_package(${ARGS})
+
+if(TARGET arrow_static)
+    list(APPEND ARROW_LIBRARIES arrow_static)
+    list(APPEND PARQUET_LIBRARIES parquet_static)
+elseif (TARGET arrow_shared)
+    list(APPEND ARROW_LIBRARIES arrow_shared)
+    list(APPEND PARQUET_LIBRARIES parquet_shared)
+endif()

externals/vcpkg/ports/arrow/vcpkg.json

@@ -0,0 +1,110 @@
+{
+  "name": "arrow",
+  "version": "8.0.0",
+  "description": "Cross-language development platform for in-memory analytics",
+  "homepage": "https://arrow.apache.org",
+  "license": "Apache-2.0",
+  "supports": "x64 | (arm64 & !windows)",
+  "dependencies": [
+    "boost-filesystem",
+    "boost-multiprecision",
+    "boost-system",
+    "brotli",
+    "bzip2",
+    "gflags",
+    "glog",
+    "lz4",
+    "openssl",
+    "re2",
+    "snappy",
+    "thrift",
+    "utf8proc",
+    {
+      "name": "vcpkg-cmake",
+      "host": true
+    },
+    {
+      "name": "vcpkg-cmake-config",
+      "host": true
+    },
+    "zlib",
+    "zstd"
+  ],
+  "default-features": [
+    "csv",
+    "filesystem",
+    "json",
+    "parquet"
+  ],
+  "features": {
+    "csv": {
+      "description": "CSV support"
+    },
+    "cuda": {
+      "description": "cuda support",
+      "dependencies": [
+        "cuda"
+      ]
+    },
+    "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",
+      "dependencies": [
+        "rapidjson"
+      ]
+    },
+    "plasma": {
+      "description": "Plasma support",
+      "supports": "!windows"
+    },
+    "s3": {
+      "description": "S3 support",
+      "dependencies": [
+        {
+          "name": "aws-sdk-cpp",
+          "default-features": false,
+          "features": [
+            "cognito-identity",
+            "config",
+            "identity-management",
+            "s3",
+            "sts",
+            "transfer"
+          ]
+        }
+      ]
+    }
+  }
+}

externals/vcpkg/ports/arrow/vs-2022-fixes.patch

@@ -0,0 +1,42 @@
+diff --git a/cpp/src/arrow/compute/light_array_test.cc b/cpp/src/arrow/compute/light_array_test.cc
+index 3f6d47803..dcc7841a0 100644
+--- a/cpp/src/arrow/compute/light_array_test.cc
++++ b/cpp/src/arrow/compute/light_array_test.cc
+@@ -18,6 +18,7 @@
+ #include "arrow/compute/light_array.h"
+ 
+ #include <gtest/gtest.h>
++#include <numeric>
+ 
+ #include "arrow/compute/exec/test_util.h"
+ #include "arrow/testing/generator.h"
+diff --git a/cpp/src/arrow/util/tracing.cc b/cpp/src/arrow/util/tracing.cc
+index b8bddcd50..126c69781 100644
+--- a/cpp/src/arrow/util/tracing.cc
++++ b/cpp/src/arrow/util/tracing.cc
+@@ -38,6 +38,7 @@ Span::Impl& Span::Set(Impl&& impl) {
+ #endif
+ 
+ // Default destructor when impl type is complete.
++Span::Span() = default;
+ Span::~Span() = default;
+ 
+ }  // namespace tracing
+diff --git a/cpp/src/arrow/util/tracing.h b/cpp/src/arrow/util/tracing.h
+index 15f7fca1e..1d9d08b33 100644
+--- a/cpp/src/arrow/util/tracing.h
++++ b/cpp/src/arrow/util/tracing.h
+@@ -35,11 +35,11 @@ class SpanImpl;
+ namespace util {
+ namespace tracing {
+ 
+-class ARROW_EXPORT Span {
++class Span {
+  public:
+   using Impl = arrow::internal::tracing::SpanImpl;
+ 
+-  Span() = default;  // Default constructor. The inner_impl is a nullptr.
++  Span();  // Default constructor. The inner_impl is a nullptr.
+   ~Span();  // Destructor. Default destructor defined in tracing.cc where impl is a
+             // complete type.
+