Using rang with package/R version inputs?

[Anouk-Bouma]

Hi,

I am working on creating a procedure to run an extensive reproducibility study, and rang is a tool that fills some major gaps.

However, I am wondering if it would be possible to specify some of the version numbers myself as a user. For example, I am trying to reproduce a paper for which the codefiles were uploaded in 2025 to OSF (so there is no version control). At the top of their script, they list the packages + version numbers and the version of R they used for their analyses. These versions are, however, much more outdated than 2025.

As a workaround, I tried using different snapshot dates in rang. But it was not possible to find the correct combination of version numbers. This is probably because the authors did not install or update all their packages at the same date, so the date is not the best way to retrieve the actual versions they were working with.

I would be very interested in an extension to the rang package in which I could input the version numbers, and ideally also input the sessionInfo() output as the input for rang to work with.

All the best!

[chainsawriot]

For this case (you have the version numbers of all packages), I would say renv (and perhaps installing packages via pak::pkg_install("cran::sna@2.4"), see https://pak.r-lib.org/reference/pak_package_sources.html) is probably a better suit for your use case.

But surely, I can imagine something like

x <- resolve(pkgs = c("sna@2.4", "schochastics/rtoot", "S4Vectors"), snapshot_date = "2022-11-30")

could be a solution, which has not been implemented yet. Although I can imagine recursive dependencies could be a problem.