GH-48010: [C++] Update bundled RE2 from 2022-06-01 to 2023-03-01

[thisisnic]

Rationale for this change

Arrow currently bundles RE2 version 2022-06-01, which fails to build on musl libc systems (Alpine Linux) due to missing #include <cstdint> in RE2's util/pcre.h. This affects R package installations from source on Alpine (#46769), C++ builds on Alpine (#43350, #41619), and causes CRAN extended musl checks to fail.

What changes are included in this PR?

Updated cpp/thirdparty/versions.txt

Are these changes tested?

Yep - CI checks

Are there any user-facing changes?

No.

• GitHub Issue: [C++] Update bundled RE2 to fix musl/Alpine Linux build failure #48010

[github-actions]

⚠️ GitHub issue #48010 has been automatically assigned in GitHub to PR creator.

[thisisnic]

Ah wait, nope, RE2 needed to be able to find Abseil.

[thisisnic]

OK, this might not be the correct solution, we get (presumably) a ton of ABI-compatibility issues here :(

[jonkeane]

What errors were you seeing with the version bumps? The ones I see in CI are strange, something is wrong, but it did look like things built successfully, which is a good sign. But maybe you were running things locally that showed this isn't a path we could pursue?

[jonkeane]

FWIW: the CI errors you got up there are both on main, I believe

[thisisnic]

@github-actions crossbow submit -g cpp

[thisisnic]

Thanks for pointing that out @jonkeane!

[github-actions]

Revision: 1becf1f

Submitted crossbow builds: ursacomputing/crossbow @ actions-07dc423666

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-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

[thisisnic]

The failing tests (ARM64 macOS 14 C++ and GLib & Ruby) are also failing on main - not related to this RE2 update. See run on main with identical substrait test failures.

[pitrou]

They haven't released since 2023?

[thisisnic]

They haven't released since 2023?

They have, but they take on Abseil as a dependency later, and my understanding is that it's a pretty big component to build just to get the RE2 library, and would slow things down for folks who don't need Abseil, and be a pain for our CRAN builds. That might be not be a reason to not bump it further though; it was just my focus at the time.

[pitrou]

They have, but they take on Abseil as a dependency later, and my understanding is that it's a pretty big component to build just to get the RE2 library, and would slow things down for folks who don't need Abseil, and be a pain for our CRAN builds.

Thanks for the explanation, I agree very much with it.

[thisisnic]

I just realised that an earlier version of this PR did bump to a version that requires Abseil which is why I have some extra changes in there to cpp/cmake_modules/ThirdpartyToolchain.cmake. I'll get rid of that.

[kou]

+1

Could you add a note why we don't use more recent RE2 (no Abseil) to cpp/thirdparty/versions.txt (as a comment) and the PR description before we merge this PR? We'll revisit it later.

[thisisnic]

+1

Could you add a note why we don't use more recent RE2 (no Abseil) to cpp/thirdparty/versions.txt (as a comment) and the PR description before we merge this PR? We'll revisit it later.

Absolutely!

[jonkeane]

Thanks for this ❤️

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 0 benchmarking runs that have been run so far on merge-commit d165954.

None of the specified runs were found on the Conbench server.

The full Conbench report has more details.