GH-48315: [C++] Use FetchContent for bundled CRC32C

[raulcd]

Rationale for this change

As a follow up of requiring a minimum CMake version >= 3.25 we discussed moving our dependencies from ExternalProject to FetchContent. This can simplify our third party dependency management.

What changes are included in this PR?

The general change is moving CRC32C from ExternalProject to FetchContent.

We will be able to clean up installation once google-cloud-cpp is also migated.

Are these changes tested?

Yes, the changes are tested locally and on CI.

Are there any user-facing changes?

No

• GitHub Issue: [C++] Use FetchContent for bundled CRC32C #48315

[raulcd]

@github-actions crossbow submit -g cpp

[github-actions]

Revision: e4ed6a0

Submitted crossbow builds: ursacomputing/crossbow @ actions-7215a8e6f9

Task	Status
example-cpp-minimal-build-static
example-cpp-minimal-build-static-system-dependency
example-cpp-tutorial
test-build-cpp-fuzz
test-conda-cpp
test-conda-cpp-valgrind
test-cuda-cpp-ubuntu-22.04-cuda-11.7.1
test-cuda-cpp-ubuntu-24.04-cuda-13.0.2
test-debian-12-cpp-amd64
test-debian-12-cpp-i386
test-fedora-42-cpp
test-ubuntu-22.04-cpp
test-ubuntu-22.04-cpp-20
test-ubuntu-22.04-cpp-bundled
test-ubuntu-22.04-cpp-emscripten
test-ubuntu-22.04-cpp-no-threading
test-ubuntu-24.04-cpp
test-ubuntu-24.04-cpp-bundled-offline
test-ubuntu-24.04-cpp-gcc-13-bundled
test-ubuntu-24.04-cpp-gcc-14
test-ubuntu-24.04-cpp-minimal-with-formats
test-ubuntu-24.04-cpp-thread-sanitizer

[raulcd]

@github-actions crossbow submit -g r

[github-actions]

Revision: e4ed6a0

Submitted crossbow builds: ursacomputing/crossbow @ actions-e9c7382c41

Task	Status
r-binary-packages
r-recheck-most
test-r-arrow-backwards-compatibility
test-r-depsource-bundled
test-r-depsource-system
test-r-dev-duckdb
test-r-devdocs
test-r-extra-packages
test-r-gcc-11
test-r-gcc-12
test-r-install-local
test-r-install-local-minsizerel
test-r-linux-as-cran
test-r-linux-rchk
test-r-linux-sanitizers
test-r-linux-valgrind
test-r-m1-san
test-r-macos-as-cran
test-r-minimal-build
test-r-offline-maximal
test-r-offline-minimal
test-r-rhub-debian-gcc-devel-lto-latest
test-r-rhub-ubuntu-gcc12-custom-ccache
test-r-rhub-ubuntu-release-latest
test-r-rocker-r-ver-latest
test-r-rstudio-r-base-4.1-focal
test-r-rstudio-r-base-4.2-focal
test-r-ubuntu-22.04
test-r-versions

[raulcd]

I have continued working on this but I am unsure we can make google-cloud-cpp use FetchContent. This is what I am currently investigating.
The problem is that google-cloud-cpp unconditionally does:
find_package(absl CONFIG REQUIRED)
https://github.com/googleapis/google-cloud-cpp/blob/d6aff236f88b608fd15e423cfe74a04c83b90882/CMakeLists.txt#L221

Abseil at that point is only configured and it tries to load the targets file due to the following on build/_deps/absl-build/abslConfig.cmake:

include ("${CMAKE_CURRENT_LIST_DIR}/abslTargets.cmake")

but the abslTargets.cmake file is only generated at install time.

@kou do you have ideas for that? I'd love to get rid of all the manual target definitions for abseil but I am unsure what can we do without patching google-cloud-cpp or keeping it as ExternalProject_Add

[raulcd]

If I apply a patch like on google-cloud-cpp I am able to bypass the problems:

--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -210,7 +210,11 @@ add_custom_target(google-cloud-cpp-protos)
 add_custom_target(doxygen-docs)
 add_custom_target(all-docfx)

-find_package(absl CONFIG REQUIRED)
+# Allow using absl from FetchContent (targets already exist)
+if(NOT TARGET absl::base)
+  find_package(absl CONFIG REQUIRED)
+endif()
+
 if (GOOGLE_CLOUD_CPP_ENABLE_GRPC)
     find_package(gRPC REQUIRED QUIET)
     # `Protobuf_PROTOC_EXECUTABLE` is pretty standard. It has been in use by
--- a/cmake/IncludeNlohmannJson.cmake
+++ b/cmake/IncludeNlohmannJson.cmake
@@ -15,6 +15,11 @@
 # ~~~

 function (find_nlohmann_json)
+    # Allow using nlohmann_json from FetchContent (target already exists)
+    if(TARGET nlohmann_json::nlohmann_json)
+        return()
+    endif()
+
     find_package(nlohmann_json CONFIG QUIET)
     if (nlohmann_json_FOUND)
         return()

Is not strictly for this PR but wanted to share the full picture as the current PR is basicaly pre-work for the google-cloud-cpp one.

[kou]

Could you try OVERRIDE_FIND_PACKAGE in FetchContent_Declare()?
https://cmake.org/cmake/help/latest/module/FetchContent.html#integrating-with-find-package

[raulcd]

Could you try OVERRIDE_FIND_PACKAGE in FetchContent_Declare()? https://cmake.org/cmake/help/latest/module/FetchContent.html#integrating-with-find-package

Thanks a lot @kou, that works as expected locally! that was much easier and better than my patch approach :)

One day I'll finally learn these various CMake options. I've learnt a lot the last couple of years, I had never touched a line of CMake prior to working on Arrow, but things like this show me there's still a lot to learn.

[kou]

Could you push a change you tried? (I think that uses OVERRIDE_FIND_PACKAGE. :-)

[raulcd]

Could you push a change you tried? (I think that uses OVERRIDE_FIND_PACKAGE. :-)

It does not seem necessary for Crc32c, at least locally, it is for Abseil and nlohmann-json. I have pushed a separate draft PR with all the google-cloud-cpp changes to exercise CI and to show you but I prefer this incremental approach so I can tackle problems individually.
The draft PR with the full changes is this one (which incorporates both Crc32c and nlohmann-json)

• GH-48317: [C++] Use FetchContent for bundled google-cloud-cpp  #48333

and the specific commit where OVERRIDE_FIND_PACKAGE is required: e24724f

[raulcd]

Sorry, after some more testing on the mentioned PR it seems it is necessary for some scenarios. I am still fixing a couple of issues on the other PR. I'll add the OVERRIDE_FIND_PACKAGE here.

[kou]

+1

Ah, sorry. I misunderstood.