GH-49764: [C++][Python] Avoid building bundled Abseil outside resolve_dependency

[raulcd]

Rationale for this change

With the bump versions for gRPC/Protobuf/Abseil and others a bug was introduced where we build Abseil outside of resolve dependency. This caused to interpret Abseil as System dependency for Arrow Flight when it was bundled and failing to find it when trying to build PyArrow.

What changes are included in this PR?

Reorder resolve_dependency abseil to be done before resolving protobuf in order to not require manually building abseil.

Are these changes tested?

Yes via CI, via archery and validated the local build that previously failed now succeeds

Are there any user-facing changes?

No

• GitHub Issue: [Python][C++] PyArrow fails to build trying to find a system abseil when Arrow C++ was built with bundled abseil #49764

[github-actions]

⚠️ GitHub issue #49764 has been automatically assigned in GitHub to PR creator.

[raulcd]

@github-actions crossbow submit verify-rc-source-python-linux-*

[github-actions]

Revision: c8be506

Submitted crossbow builds: ursacomputing/crossbow @ actions-71cba4b15c

Task	Status
verify-rc-source-python-linux-almalinux-10-amd64
verify-rc-source-python-linux-conda-latest-amd64
verify-rc-source-python-linux-ubuntu-22.04-amd64
verify-rc-source-python-linux-ubuntu-24.04-amd64

[raulcd]

@kou this is just re-ordering resolve dependencies.
It avoids some wrong assumptions down the line due to an early build Abseil from Protobuf, andFlight setting flags as if Abseil is a system dependency instead of being bundled.
As is just moving things around (plus a couple more checks due to Abseil being built for Protobuf) I plan to merge this if no concerns are raised.

[kou]

+1

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 0 benchmarking runs that have been run so far on merge-commit 34414b5.

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

The full Conbench report has more details.