GH-45508: [CI][R] Remove Ubuntu version from sanitizer jobs

[kou]

Rationale for this change

ubuntu-r-sanitizer and ubuntu-r-valgrind use wch1/r-debug as their base image.

So we can't control Ubuntu versions for them.

What changes are included in this PR?

Remove Ubuntu version from their configurations.

Are these changes tested?

Yes.

Are there any user-facing changes?

No.

• GitHub Issue: [CI][R] Remove Ubuntu versions from sanitizer jobs #45508

[github-actions]

⚠️ GitHub issue #45508 has been automatically assigned in GitHub to PR creator.

[kou]

@github-actions crossbow submit test-ubuntu-r-sanitizer test-r-linux-valgrind

[pitrou]

Why not use the more modern Ubuntu 24.04?

[kou]

Ah, wait. I was wrong.

These jobs already use Ubuntu 22.04 when update UBUNTU in .env.

Only labels and comments use 20.04.

Let's use UBUNTU in .env instead of 24.04 for easy to maintain. (If we use UBUNTU in .env, we don't need to update Ubuntu versions for them separately.)

[kou]

No, I was wrong...

ubuntu-r-sanitizer uses wch1/r-debug as its base image:

arrow/docker-compose.yml

Line 1618 in f8ba543

base: wch1/r-debug:latest

It uses Ubuntu 22.04: https://github.com/wch/r-debug/blob/13dd42827354c94f9cb4f0bec69c9164c5a700cd/r-devel/Dockerfile#L9

[kou]

ubuntu-r-valgrind uses wch1/r-debug too.

Anyway, we can't control Ubuntu version for them.

[kou]

@github-actions crossbow submit test-r-linux-sanitizer test-r-linux-valgrind

[github-actions]

Unable to match any tasks for `test-ubuntu-r-sanitizer`
The Archery job run can be found at: https://github.com/apache/arrow/actions/runs/13298990192

[jonkeane]

Thanks for this (and all of the other recent cleanup on this)

[kou]

@github-actions crossbow submit test-r-linux-sanitizer test-r-linux-valgrind

[github-actions]

Revision: 700a3df

Submitted crossbow builds: ursacomputing/crossbow @ actions-fbe0d04a54

Task	Status
test-r-linux-sanitizer
test-r-linux-valgrind

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit 4e65a94.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 1 possible false positive for unstable benchmarks that are known to sometimes produce them.