[arrow] Update to 5.0.0 (#19229)

* Bump version to 5.0.0

* Update version files

* Update patches

* Update version files

* Fix patch

* Update version files

* Add newline at end of all.patch

* Update version files

* Unpin utf8proc

* Update version files

* Expect tensorflow-cc build fail

* Remove tensorflow-cc baseline failure

6 files changed, 77 insertions, 69 deletions

diff --git a/ports/arrow/all.patch b/ports/arrow/all.patch
index 180fafc58..a9b3a2628 100644
--- a/ports/arrow/all.patch
+++ b/ports/arrow/all.patch
@@ -1,8 +1,8 @@
 diff --git a/cpp/cmake_modules/BuildUtils.cmake b/cpp/cmake_modules/BuildUtils.cmake
-index 2fd897b5d..b6118ad4f 100644
+index cd8290d1b..12c52c184 100644
 --- a/cpp/cmake_modules/BuildUtils.cmake
 +++ b/cpp/cmake_modules/BuildUtils.cmake
-@@ -440,7 +440,7 @@ function(ADD_ARROW_LIB LIB_NAME)
+@@ -427,7 +427,7 @@ function(ADD_ARROW_LIB LIB_NAME)
        target_include_directories(${LIB_NAME}_static PRIVATE ${ARG_PRIVATE_INCLUDES})
      endif()
  
diff --git a/ports/arrow/fix-dependencies.patch b/ports/arrow/fix-dependencies.patch
index 65dcee334..3e22d1801 100644
--- a/ports/arrow/fix-dependencies.patch
+++ b/ports/arrow/fix-dependencies.patch
@@ -1,8 +1,8 @@
 diff --git a/cpp/CMakeLists.txt b/cpp/CMakeLists.txt
-index b7a1dae..2669b15 100644
+index b5a39bf33..af1e16a9b 100644
 --- a/cpp/CMakeLists.txt
 +++ b/cpp/CMakeLists.txt
-@@ -669,7 +669,7 @@ endif()
+@@ -668,7 +668,7 @@ endif()
  
  if(ARROW_WITH_BROTLI)
    # Order is important for static linking
@@ -11,7 +11,7 @@ index b7a1dae..2669b15 100644
    list(APPEND ARROW_LINK_LIBS ${ARROW_BROTLI_LIBS})
    list(APPEND ARROW_STATIC_LINK_LIBS ${ARROW_BROTLI_LIBS})
    if(Brotli_SOURCE STREQUAL "SYSTEM")
-@@ -685,9 +685,9 @@ if(ARROW_WITH_BZ2)
+@@ -684,9 +684,9 @@ if(ARROW_WITH_BZ2)
  endif()
  
  if(ARROW_WITH_LZ4)
@@ -23,7 +23,7 @@ index b7a1dae..2669b15 100644
    endif()
  endif()
  
-@@ -736,10 +736,10 @@ if(ARROW_S3)
+@@ -735,10 +735,10 @@ if(ARROW_S3)
  endif()
  
  if(ARROW_WITH_UTF8PROC)
@@ -38,7 +38,7 @@ index b7a1dae..2669b15 100644
  endif()
  
 diff --git a/cpp/cmake_modules/ThirdpartyToolchain.cmake b/cpp/cmake_modules/ThirdpartyToolchain.cmake
-index 83ea3aa..d592fc3 100644
+index e6852d9c2..fc0973ffa 100644
 --- a/cpp/cmake_modules/ThirdpartyToolchain.cmake
 +++ b/cpp/cmake_modules/ThirdpartyToolchain.cmake
 @@ -53,7 +53,7 @@ set(ARROW_THIRDPARTY_DEPENDENCIES
@@ -67,22 +67,22 @@ index 83ea3aa..d592fc3 100644
      xsimd
      ZLIB
      zstd)
-@@ -956,7 +956,7 @@ macro(build_snappy)
+@@ -945,7 +945,7 @@ macro(build_snappy)
  endmacro()
  
  if(ARROW_WITH_SNAPPY)
--  resolve_dependency(Snappy)
-+  resolve_dependency(Snappy USE_CONFIG TRUE)
-   # TODO: Don't use global includes but rather target_include_directories
-   get_target_property(SNAPPY_INCLUDE_DIRS Snappy::snappy INTERFACE_INCLUDE_DIRECTORIES)
-   include_directories(SYSTEM ${SNAPPY_INCLUDE_DIRS})
-@@ -1021,10 +1021,16 @@ macro(build_brotli)
+-  resolve_dependency(Snappy PC_PACKAGE_NAMES snappy)
++  resolve_dependency(Snappy USE_CONFIG TRUE PC_PACKAGE_NAMES snappy)
+   if(${Snappy_SOURCE} STREQUAL "SYSTEM" AND NOT snappy_PC_FOUND)
+     get_target_property(SNAPPY_LIB Snappy::snappy IMPORTED_LOCATION)
+     string(APPEND ARROW_PC_LIBS_PRIVATE " ${SNAPPY_LIB}")
+@@ -1014,10 +1014,16 @@ macro(build_brotli)
  endmacro()
  
  if(ARROW_WITH_BROTLI)
--  resolve_dependency(Brotli)
+-  resolve_dependency(Brotli PC_PACKAGE_NAMES libbrotlidec libbrotlienc)
 +  set(unofficial-brotli_SOURCE "SYSTEM")
-+  resolve_dependency(unofficial-brotli USE_CONFIG TRUE)
++  resolve_dependency(unofficial-brotli USE_CONFIG TRUE PC_PACKAGE_NAMES libbrotlidec libbrotlienc)
    # TODO: Don't use global includes but rather target_include_directories
 -  get_target_property(BROTLI_INCLUDE_DIR Brotli::brotlicommon
 -                      INTERFACE_INCLUDE_DIRECTORIES)
@@ -96,16 +96,16 @@ index 83ea3aa..d592fc3 100644
    include_directories(SYSTEM ${BROTLI_INCLUDE_DIR})
  endif()
  
-@@ -1141,7 +1147,7 @@ macro(build_glog)
+@@ -1133,7 +1139,7 @@ macro(build_glog)
  endmacro()
  
  if(ARROW_USE_GLOG)
--  resolve_dependency(GLOG)
-+  resolve_dependency(glog USE_CONFIG TRUE)
+-  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})
-@@ -1214,8 +1220,7 @@ endmacro()
+@@ -1206,8 +1212,7 @@ endmacro()
  if(ARROW_NEED_GFLAGS)
    set(ARROW_GFLAGS_REQUIRED_VERSION "2.1.0")
    resolve_dependency(gflags
@@ -115,24 +115,23 @@ index 83ea3aa..d592fc3 100644
                       REQUIRED_VERSION
                       ${ARROW_GFLAGS_REQUIRED_VERSION}
                       IS_RUNTIME_DEPENDENCY
-@@ -1316,7 +1321,7 @@ if(ARROW_WITH_THRIFT)
-   # to build Boost, so don't look again if already found.
+@@ -1309,6 +1314,7 @@ if(ARROW_WITH_THRIFT)
    if(NOT Thrift_FOUND AND NOT THRIFT_FOUND)
      # Thrift c++ code generated by 0.13 requires 0.11 or greater
--    resolve_dependency(Thrift REQUIRED_VERSION 0.11.0)
-+    resolve_dependency(Thrift USE_CONFIG TRUE REQUIRED_VERSION 0.11.0)
-   endif()
-   # TODO: Don't use global includes but rather target_include_directories
-   include_directories(SYSTEM ${THRIFT_INCLUDE_DIR})
-@@ -1416,13 +1421,17 @@ if(ARROW_WITH_PROTOBUF)
-   else()
+     resolve_dependency(Thrift
++                       USE_CONFIG TRUE
+                        REQUIRED_VERSION
+                        0.11.0
+                        PC_PACKAGE_NAMES
+@@ -1413,6 +1419,7 @@ if(ARROW_WITH_PROTOBUF)
      set(ARROW_PROTOBUF_REQUIRED_VERSION "2.6.1")
    endif()
--  resolve_dependency(Protobuf REQUIRED_VERSION ${ARROW_PROTOBUF_REQUIRED_VERSION})
-+  resolve_dependency(Protobuf USE_CONFIG TRUE REQUIRED_VERSION ${ARROW_PROTOBUF_REQUIRED_VERSION})
- 
-   if(ARROW_PROTOBUF_USE_SHARED AND MSVC_TOOLCHAIN)
-     add_definitions(-DPROTOBUF_USE_DLLS)
+   resolve_dependency(Protobuf
++                     USE_CONFIG TRUE
+                      REQUIRED_VERSION
+                      ${ARROW_PROTOBUF_REQUIRED_VERSION}
+                      PC_PACKAGE_NAMES
+@@ -1423,6 +1430,10 @@ if(ARROW_WITH_PROTOBUF)
    endif()
  
    # TODO: Don't use global includes but rather target_include_directories
@@ -143,7 +142,7 @@ index 83ea3aa..d592fc3 100644
    include_directories(SYSTEM ${PROTOBUF_INCLUDE_DIR})
  
    if(TARGET arrow::protobuf::libprotobuf)
-@@ -1471,12 +1480,18 @@ if(ARROW_WITH_PROTOBUF)
+@@ -1471,12 +1482,18 @@ if(ARROW_WITH_PROTOBUF)
  
    # Log protobuf paths as we often see issues with mixed sources for
    # the libraries and protoc.
@@ -162,7 +161,7 @@ index 83ea3aa..d592fc3 100644
    message(STATUS "Found libprotobuf: ${PROTOBUF_LIBRARY}")
    message(STATUS "Found protobuf headers: ${PROTOBUF_INCLUDE_DIR}")
  endif()
-@@ -1904,7 +1919,7 @@ endmacro()
+@@ -1889,7 +1906,7 @@ endmacro()
  if(ARROW_WITH_RAPIDJSON)
    set(ARROW_RAPIDJSON_REQUIRED_VERSION "1.1.0")
    resolve_dependency(RapidJSON
@@ -171,13 +170,13 @@ index 83ea3aa..d592fc3 100644
                       TRUE
                       REQUIRED_VERSION
                       ${ARROW_RAPIDJSON_REQUIRED_VERSION}
-@@ -2047,10 +2062,11 @@ macro(build_lz4)
+@@ -2024,10 +2041,11 @@ macro(build_lz4)
  endmacro()
  
  if(ARROW_WITH_LZ4)
--  resolve_dependency(Lz4)
+-  resolve_dependency(Lz4 PC_PACKAGE_NAMES liblz4)
 +  set(lz4_SOURCE "SYSTEM")
-+  resolve_dependency(lz4 USE_CONFIG TRUE)
++  resolve_dependency(lz4 USE_CONFIG TRUE PC_PACKAGE_NAMES liblz4)
  
    # TODO: Don't use global includes but rather target_include_directories
 -  get_target_property(LZ4_INCLUDE_DIR LZ4::lz4 INTERFACE_INCLUDE_DIRECTORIES)
@@ -185,23 +184,28 @@ index 83ea3aa..d592fc3 100644
    include_directories(SYSTEM ${LZ4_INCLUDE_DIR})
  endif()
  
-@@ -2111,7 +2127,7 @@ macro(build_zstd)
- endmacro()
- 
+@@ -2090,7 +2108,7 @@ endmacro()
  if(ARROW_WITH_ZSTD)
--  resolve_dependency(zstd)
-+  resolve_dependency(zstd USE_CONFIG TRUE)
- 
-   if(TARGET zstd::libzstd)
-     set(ARROW_ZSTD_LIBZSTD zstd::libzstd)
-@@ -2273,13 +2289,13 @@ macro(build_utf8proc)
+   # ARROW-13384: ZSTD_minCLevel was added in v1.4.0, required by ARROW-13091
+   resolve_dependency(zstd
+-                     PC_PACKAGE_NAMES
++                     USE_CONFIG TRUE PC_PACKAGE_NAMES
+                      libzstd
+                      REQUIRED_VERSION
+                      1.4.0)
+@@ -2262,9 +2280,8 @@ macro(build_utf8proc)
  endmacro()
  
  if(ARROW_WITH_UTF8PROC)
--  resolve_dependency(utf8proc)
-+  resolve_dependency(unofficial-utf8proc USE_CONFIG TRUE)
+-  resolve_dependency(utf8proc
+-                     REQUIRED_VERSION
+-                     "2.2.0"
++  resolve_dependency(unofficial-utf8proc
++                     USE_CONFIG TRUE
+                      PC_PACKAGE_NAMES
+                      libutf8proc)
  
-   add_definitions(-DARROW_WITH_UTF8PROC)
+@@ -2272,7 +2289,7 @@ if(ARROW_WITH_UTF8PROC)
  
    # TODO: Don't use global definitions but rather
    # target_compile_definitions or target_link_libraries
@@ -210,7 +214,7 @@ index 83ea3aa..d592fc3 100644
                        INTERFACE_COMPILER_DEFINITIONS)
    if(UTF8PROC_COMPILER_DEFINITIONS)
      add_definitions(-D${UTF8PROC_COMPILER_DEFINITIONS})
-@@ -2287,7 +2303,7 @@ if(ARROW_WITH_UTF8PROC)
+@@ -2280,7 +2297,7 @@ if(ARROW_WITH_UTF8PROC)
  
    # TODO: Don't use global includes but rather
    # target_include_directories or target_link_libraries
@@ -219,16 +223,16 @@ index 83ea3aa..d592fc3 100644
                        INTERFACE_INCLUDE_DIRECTORIES)
    include_directories(SYSTEM ${UTF8PROC_INCLUDE_DIR})
  endif()
-@@ -2342,7 +2358,7 @@ endmacro()
- # Dependencies for Arrow Flight RPC
+@@ -2335,7 +2352,7 @@ endmacro()
  
  macro(build_grpc)
--  resolve_dependency(c-ares HAVE_ALT TRUE)
-+  resolve_dependency(c-ares HAVE_ALT TRUE USE_CONFIG TRUE)
-   # TODO: Don't use global includes but rather target_include_directories
-   get_target_property(c-ares_INCLUDE_DIR c-ares::cares INTERFACE_INCLUDE_DIRECTORIES)
-   include_directories(SYSTEM ${c-ares_INCLUDE_DIR})
-@@ -2614,7 +2630,7 @@ endmacro()
+   resolve_dependency(c-ares
+-                     HAVE_ALT
++                     USE_CONFIG
+                      TRUE
+                      PC_PACKAGE_NAMES
+                      libcares)
+@@ -2600,7 +2617,7 @@ endmacro()
  if(ARROW_WITH_GRPC)
    set(ARROW_GRPC_REQUIRED_VERSION "1.17.0")
    resolve_dependency(gRPC
@@ -236,8 +240,8 @@ index 83ea3aa..d592fc3 100644
 +                     USE_CONFIG
                       TRUE
                       REQUIRED_VERSION
-                      ${ARROW_GRPC_REQUIRED_VERSION})
-@@ -2622,6 +2638,10 @@ if(ARROW_WITH_GRPC)
+                      ${ARROW_GRPC_REQUIRED_VERSION}
+@@ -2610,6 +2627,10 @@ 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})
@@ -249,7 +253,7 @@ index 83ea3aa..d592fc3 100644
    if(GRPC_VENDORED)
      set(GRPCPP_PP_INCLUDE TRUE)
 diff --git a/cpp/src/arrow/adapters/orc/CMakeLists.txt b/cpp/src/arrow/adapters/orc/CMakeLists.txt
-index 516196c..ebc4881 100644
+index ca901b07d..4512a175f 100644
 --- a/cpp/src/arrow/adapters/orc/CMakeLists.txt
 +++ b/cpp/src/arrow/adapters/orc/CMakeLists.txt
 @@ -29,7 +29,7 @@ set(ORC_MIN_TEST_LIBS
diff --git a/ports/arrow/portfile.cmake b/ports/arrow/portfile.cmake
index bb7d857fb..221b3e0d9 100644
--- a/ports/arrow/portfile.cmake
+++ b/ports/arrow/portfile.cmake
@@ -3,8 +3,8 @@ vcpkg_fail_port_install(ON_ARCH "x86" "arm" "arm64")
 vcpkg_from_github(
     OUT_SOURCE_PATH SOURCE_PATH
     REPO apache/arrow
-    REF apache-arrow-4.0.0
-    SHA512 4697a32004d02a519b8a8e899ed3cd981ae3485e6d34071436051080d6c84e25ad0bc568b3e52effe0a9204756da3d6e560a2037df06d2730dccd19c6b4c8027
+    REF apache-arrow-5.0.0
+    SHA512 68f4377f654423e7ea47c8c0170ddb030d0b020b936ec435854e216392e6515d870287f410d0acd48b36dcfec61be4c4a95794857f1d91c66745f3c6ed748034
     HEAD_REF master
     PATCHES
         all.patch
diff --git a/ports/arrow/vcpkg.json b/ports/arrow/vcpkg.json
index c15f01450..98b768494 100644
--- a/ports/arrow/vcpkg.json
+++ b/ports/arrow/vcpkg.json
@@ -1,7 +1,6 @@
 {
   "name": "arrow",
-  "version": "4.0.0",
-  "port-version": 1,
+  "version": "5.0.0",
   "description": "Cross-language development platform for in-memory analytics",
   "homepage": "https://arrow.apache.org",
   "supports": "x64",
diff --git a/versions/a-/arrow.json b/versions/a-/arrow.json
index d4ea77e4a..904f3430c 100644
--- a/versions/a-/arrow.json
+++ b/versions/a-/arrow.json
@@ -1,6 +1,11 @@
 {
   "versions": [
     {
+      "git-tree": "bf5c2b42bb8c4efa09d69a4ce6e75ae3ace45f02",
+      "version": "5.0.0",
+      "port-version": 0
+    },
+    {
       "git-tree": "09e65930903c5dad92250b8f3f351f06f77f8e30",
       "version": "4.0.0",
       "port-version": 1
diff --git a/versions/baseline.json b/versions/baseline.json
index c6742dd55..4539f1104 100644
--- a/versions/baseline.json
+++ b/versions/baseline.json
@@ -165,8 +165,8 @@
       "port-version": 3
     },
     "arrow": {
-      "baseline": "4.0.0",
-      "port-version": 1
+      "baseline": "5.0.0",
+      "port-version": 0
     },
     "ashes": {
       "baseline": "2021-06-18",