GH-44114: [R] Add Rocky and opensuse to the allowlist for libarrow binaries

[jonkeane]

Rationale for this change

Add two distros not checked on CRAN to our allow list for easier binary installs

What changes are included in this PR?

Added to distro names

Are these changes tested?

These configurations are tested elsewhere, the real test will come later if CRAN somehow runs these covertly

Are there any user-facing changes?

Faster, more reliably installs without intervention

• GitHub Issue: [R] Add Rocky and opensuse to the allowlist for libarrow binaries #44114

[github-actions]

⚠️ GitHub issue #44114 has been automatically assigned in GitHub to PR creator.

[assignUser]

Is there a reason to be specific with -leap? Or do we want just opensuse to be ok?

[jonkeane]

Is there a reason to be specific with -leap? Or do we want just opensuse to be ok?

I discovered this when chatting with someone from the Posit Package Manager, and that's the string they got when they ran the distro code. But this might be an issue with distro, or maybe opensuse has different flavors? Or also possible that the Posit Package Manager folks are using a slightly non-standard build of opensuse. Do you know of any references for which to look for? I'm happy to add it, but I think we would need both to catch the system that they seem to be using.

[assignUser]

I see but with any(grepl(paste(allowlist, collapse = "|"), os)) opensuse should cover opensuse-leap and any future versions if I'm not mistaken (not at a pc right now to test it^^)

[nealrichardson]

I don't recall the history on this, but the local copy of the allowlist should be present here, right? Can we fall back to readLines("nixlibs-allowlist.txt") (or whatever path it is relative to cwd)? That way we only have to track this in one place.

Also we should update the comment on the line above.

[jonkeane]

Ah, of course

[jonkeane]

but with any(grepl(paste(allowlist, collapse = "|"), os)) opensuse should cover opensuse-leap and any future versions if I'm not mistaken

Ah, right right. Ok I've changed it to opensuse. It would be good to have confirmation of what the possible values are there, in case we need to add something like open-suse or opensus3 or something. But this should be good for the issue at hand.

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit 0707c04.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 3 possible false positives for unstable benchmarks that are known to sometimes produce them.