GH-46137: [C++] Replace grpc-cpp conda package with libgrpc

[WillAyd]

Rationale for this change

The grpc-cpp conda package appears abandoned and is artificially pinning other dependencies to older versions (ex: orc). The more recent libgrpc conda package appears to be where the development is going forward

What changes are included in this PR?

Replaced grpc-cpp with libgrpc

Are these changes tested?

Yes

Are there any user-facing changes?

No

• GitHub Issue: [C++][Packaging] Remove pin for grpc-cpp in conda_env_cpp.txt #46137

[github-actions]

⚠️ GitHub issue #47605 has been automatically assigned in GitHub to PR creator.

[WillAyd]

In CI orc is pinned to 1.9.0. This new pin is required due to some upstream bugs with the CMake configuration - see apache/orc#2408

[WillAyd]

The Go failure seems related although I'm not sure I can see anything in the logs that points to the issue. Maybe @zeroshade has an idea?

https://github.com/apache/arrow/actions/runs/17863409077/job/50799734977?pr=47606#step:13:1204

[h-vetinari]

The grpc-cpp conda package appears abandoned and is artificially pinning other dependencies to older versions (ex: orc). The more recent libgrpc conda package appears to be where the development is going forward

Indeed the package was renamed (similar to the conda package for arrow going from arrow-cpp to libarrow) and we eventually dropped the compatibility wrapper for the old naming, which is why the last available builds will pull in very old dependencies.

[WillAyd]

It looks like the integration failure here was reported by @kou upstream in #47673

FWIW it has happened with every push so far on this branch, but maybe that has been bad luck

[WillAyd]

OK this is all green - @kou any thoughts on this?

[github-actions]

⚠️ GitHub issue #46137 has been automatically assigned in GitHub to PR creator.

[kou]

+1

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 2 benchmarking runs that have been run so far on merge-commit 4b55ea6.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 17 possible false positives for unstable benchmarks that are known to sometimes produce them.