GH-44096: [C++] Don't use Boost.Process with Emscripten

[kou]

Rationale for this change

Boost.Process doesn't work with Emscripten. So we can't build arrow::util::Process.

What changes are included in this PR?

Don't use Boost.Process with Emscripten. arrow::util::Process returns Status::NotImplemented with Emscripten.

Are these changes tested?

Yes.

Are there any user-facing changes?

No.

• GitHub Issue: [C++] Can't build arrow::util::Process with Emscripten #44096

[kou]

@github-actions crossbow submit *-emscripten

[github-actions]

⚠️ GitHub issue #44096 has been automatically assigned in GitHub to PR creator.

[github-actions]

Revision: ad112ad

Submitted crossbow builds: ursacomputing/crossbow @ actions-0be59ac65c

Task	Status
test-conda-python-emscripten
test-ubuntu-22.04-cpp-emscripten

[kou]

+1

This test-conda-python-emscripten failure is unrelated:

https://github.com/ursacomputing/crossbow/actions/runs/10842738518/job/30088799844#step:7:8843

==================================== ERRORS ====================================
____________ ERROR at setup of test_column_chunk_key_value_metadata ____________

    @pytest.fixture(scope='module')
    def parquet_test_datadir():
        result = os.environ.get('PARQUET_TEST_DATA')
        if not result:
>           raise RuntimeError('Please point the PARQUET_TEST_DATA environment '
                               'variable to the test data directory')
E           RuntimeError: Please point the PARQUET_TEST_DATA environment variable to the test data directory

/lib/python3.12/site-packages/pyarrow/tests/parquet/conftest.py:35: RuntimeError

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit 62060ab.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 10 possible false positives for unstable benchmarks that are known to sometimes produce them.