arrow: add version 9 and more variants (#32701)

* utf8proc: add version 2.7.0 and shared variant
* xsimd: add version 8.1.0
* arrow: add version 9.0.0 and more variants

1 files changed, 97 insertions, 44 deletions

diff --git a/var/spack/repos/builtin/packages/arrow/package.py b/var/spack/repos/builtin/packages/arrow/package.py
index 7870a7258a..c9bc1b7b66 100644
--- a/var/spack/repos/builtin/packages/arrow/package.py
+++ b/var/spack/repos/builtin/packages/arrow/package.py
@@ -2,8 +2,8 @@
 # Spack Project Developers. See the top-level COPYRIGHT file for details.
 #
 # SPDX-License-Identifier: (Apache-2.0 OR MIT)
+
 from spack.package import *
-from spack.pkg.builtin.boost import Boost
 
 
 class Arrow(CMakePackage, CudaPackage):
@@ -15,6 +15,7 @@ class Arrow(CMakePackage, CudaPackage):
     homepage = "https://arrow.apache.org"
     url = "https://github.com/apache/arrow/archive/apache-arrow-0.9.0.tar.gz"
 
+    version("9.0.0", sha256="bb187b4b0af8dcc027fffed3700a7b891c9f76c9b63ad8925b4afb8257a2bb1b")
     version("8.0.0", sha256="19ece12de48e51ce4287d2dee00dc358fbc5ff02f41629d16076f77b8579e272")
     version("7.0.0", sha256="57e13c62f27b710e1de54fd30faed612aefa22aa41fa2c0c3bacd204dd18a8f3")
     version("4.0.1", sha256="79d3e807df4a179cfab1e7a1ab5f79d95f7b72ac2c33aba030febd125d77eb3b")
@@ -28,32 +29,67 @@ class Arrow(CMakePackage, CudaPackage):
     version("0.9.0", sha256="65f89a3910b6df02ac71e4d4283db9b02c5b3f1e627346c7b6a5982ae994af91")
     version("0.8.0", sha256="c61a60c298c30546fc0b418a35be66ef330fb81b06c49928acca7f1a34671d54")
 
-    depends_on("boost@1.60:")
-
-    # TODO: replace this with an explicit list of components of Boost,
-    # for instance depends_on('boost +filesystem')
-    # See https://github.com/spack/spack/pull/22303 for reference
-    depends_on(Boost.with_default_variants)
+    depends_on("boost@1.60: +filesystem +system")
     depends_on("cmake@3.2.0:", type="build")
-    depends_on("flatbuffers build_type=Release")  # only Release contains flatc
-    depends_on("python", when="+python")
+    depends_on("flatbuffers")
+    depends_on("llvm@:11 +clang", when="+gandiva @:3", type="build")
+    depends_on("llvm@:12 +clang", when="+gandiva @:4", type="build")
+    depends_on("llvm@:13 +clang", when="+gandiva @:7", type="build")
+    depends_on("llvm@:14 +clang", when="+gandiva @8:", type="build")
+    depends_on("lz4", when="+lz4")
+    depends_on("ninja", type="build")
+    depends_on("openssl", when="+gandiva @6.0.0:")
+    depends_on("openssl", when="@4.0.0:")
+    depends_on("orc", when="+orc")
+    depends_on("protobuf", when="+gandiva")
     depends_on("py-numpy", when="+python")
+    depends_on("python", when="+python")
     depends_on("rapidjson")
-    depends_on("snappy~shared")
-    depends_on("zlib+pic")
-    depends_on("zstd")
+    depends_on("re2+shared", when="+compute")
+    depends_on("re2+shared", when="+gandiva")
+    depends_on("snappy~shared", when="+snappy @9:")
+    depends_on("snappy~shared", when="@8:")
     depends_on("thrift+pic", when="+parquet")
-    depends_on("orc", when="+orc")
-
+    depends_on("utf8proc@2.7.0: +shared", when="+compute")
+    depends_on("utf8proc@2.7.0: +shared", when="+gandiva")
+    depends_on("xsimd@8.1.0:", when="@9.0.0:")
+    depends_on("zlib+pic", when="+zlib @9:")
+    depends_on("zlib+pic", when="@:8")
+    depends_on("zstd", when="+zstd @9:")
+    depends_on("zstd", when="@:8")
+
+    variant("brotli", default=False, description="Build support for Brotli compression")
     variant(
         "build_type",
         default="Release",
         description="CMake build type",
         values=("Debug", "FastDebug", "Release"),
     )
-    variant("python", default=False, description="Build Python interface")
+    variant(
+        "compute", default=False, description="Computational kernel functions and other support"
+    )
+    variant("gandiva", default=False, description="Build Gandiva support")
+    variant(
+        "glog",
+        default=False,
+        description="Build libraries with glog support for pluggable logging",
+    )
+    variant(
+        "hdfs",
+        default=False,
+        description="Integration with libhdfs for accessing the Hadoop Filesystem",
+    )
+    variant("ipc", default=True, description="Build the Arrow IPC extensions")
+    variant("jemalloc", default=False, description="Build the Arrow jemalloc-based allocator")
+    variant("lz4", default=False, description="Build support for lz4 compression")
+    variant("orc", default=False, description="Build integration with Apache ORC")
     variant("parquet", default=False, description="Build Parquet interface")
-    variant("orc", default=False, description="Build ORC support")
+    variant("python", default=False, description="Build Python interface")
+    variant("shared", default=True, description="Build shared libs")
+    variant("snappy", default=False, description="Build support for Snappy compression")
+    variant("tensorflow", default=False, description="Build Arrow with TensorFlow support enabled")
+    variant("zlib", default=False, description="Build support for zlib (gzip) compression")
+    variant("zstd", default=False, description="Build support for ZSTD compression")
 
     root_cmakelists_dir = "cpp"
 
@@ -63,37 +99,54 @@ class Arrow(CMakePackage, CudaPackage):
             r"(include_directories\()SYSTEM ", r"\1", "cpp/cmake_modules/ThirdpartyToolchain.cmake"
         )
 
-    def cmake_args(self):
-        args = [
-            "-DARROW_USE_SSE=ON",
-            "-DARROW_BUILD_SHARED=ON",
-            "-DARROW_BUILD_STATIC=OFF",
-            "-DARROW_BUILD_TESTS=OFF",
-            "-DARROW_WITH_BROTLI=OFF",
-            "-DARROW_WITH_LZ4=OFF",
-        ]
-
-        if self.spec.satisfies("+cuda"):
-            args.append("-DARROW_CUDA:BOOL=ON")
-        else:
-            args.append("-DARROW_CUDA:BOOL=OFF")
+        filter_file(
+            r'set\(ARROW_LLVM_VERSIONS "10" "9" "8" "7"\)',
+            'set(ARROW_LLVM_VERSIONS "11" "10" "9" "8" "7")',
+            "cpp/CMakeLists.txt",
+            when="@:2.0.0",
+        )
 
-        if self.spec.satisfies("+python"):
-            args.append("-DARROW_PYTHON:BOOL=ON")
-        else:
-            args.append("-DARROW_PYTHON:BOOL=OFF")
+        filter_file(
+            r"#include <llvm/Support/DynamicLibrary\.h>",
+            r"#include <llvm/Support/DynamicLibrary.h>" + "\n" + r"#include <llvm/Support/Host.h>",
+            "cpp/src/gandiva/engine.cc",
+            when="@2.0.0",
+        )
 
-        if self.spec.satisfies("+parquet"):
-            args.append("-DARROW_PARQUET:BOOL=ON")
-        else:
-            args.append("-DARROW_PARQUET:BOOL=OFF")
+    def cmake_args(self):
+        args = ["-DARROW_DEPENDENCY_SOURCE=SYSTEM", "-DARROW_NO_DEPRECATED_API=ON"]
 
-        if self.spec.satisfies("+orc"):
-            args.append("-DARROW_ORC:BOOL=ON")
+        if self.spec.satisfies("+shared"):
+            args.append(self.define("BUILD_SHARED", "ON"))
         else:
-            args.append("-DARROW_ORC:BOOL=OFF")
+            args.append(self.define("BUILD_SHARED", "OFF"))
+            args.append(self.define("BUILD_STATIC", "ON"))
+
+        if self.spec.satisfies("@:0.11.99"):
+            # ARROW_USE_SSE was removed in 0.12
+            # see https://issues.apache.org/jira/browse/ARROW-3844
+            args.append(self.define("ARROW_USE_SSE", "ON"))
+
+        args.append(self.define_from_variant("ARROW_COMPUTE", "compute"))
+        args.append(self.define_from_variant("ARROW_CUDA", "cuda"))
+        args.append(self.define_from_variant("ARROW_GANDIVA", "gandiva"))
+        args.append(self.define_from_variant("ARROW_GLOG", "glog"))
+        args.append(self.define_from_variant("ARROW_HDFS", "hdfs"))
+        args.append(self.define_from_variant("ARROW_IPC", "ipc"))
+        args.append(self.define_from_variant("ARROW_JEMALLOC", "jemalloc"))
+        args.append(self.define_from_variant("ARROW_ORC", "orc"))
+        args.append(self.define_from_variant("ARROW_PARQUET", "parquet"))
+        args.append(self.define_from_variant("ARROW_PYTHON", "python"))
+        args.append(self.define_from_variant("ARROW_TENSORFLOW", "tensorflow"))
+        args.append(self.define_from_variant("ARROW_WITH_BROTLI", "brotli"))
+        args.append(self.define_from_variant("ARROW_WITH_LZ4", "lz4"))
+        args.append(self.define_from_variant("ARROW_WITH_SNAPPY", "snappy"))
+        args.append(self.define_from_variant("ARROW_WITH_ZLIB", "zlib"))
+        args.append(self.define_from_variant("ARROW_WITH_ZSTD", "zstd"))
+
+        with when("@:8"):
+            for dep in ("flatbuffers", "rapidjson", "snappy", "zlib", "zstd"):
+                args.append("-D{0}_HOME={1}".format(dep.upper(), self.spec[dep].prefix))
+            args.append("-DZLIB_LIBRARIES={0}".format(self.spec["zlib"].libs))
 
-        for dep in ("flatbuffers", "rapidjson", "snappy", "zlib", "zstd"):
-            args.append("-D{0}_HOME={1}".format(dep.upper(), self.spec[dep].prefix))
-        args.append("-DZLIB_LIBRARIES={0}".format(self.spec["zlib"].libs))
         return args