session_info() doesn't output the `parallel` package version

[lcolladotor]

Hi,

As reported by Martin Morgan here, devtools::session_info() doesn't output the base package version, which is information he needed in the particular bug report linked previously.

Cheers,
Leo

[BrianDiggs]

utils::sessionInfo() also does not report the version number for base, or for any of the base packages (base, compiler, datasets, grDevices, graphics, grid, methods, parallel, splines, stats, stats4, tcltk, tools, and utils). That is because these packages are version locked with the version of R; they have the same version as the running version of R and updates are only released when a new version of R is released.

Given that, it is reasonable that devtools::session_info() also does not report the version number.

[lcolladotor]

Aye, Dan Tenenbaum made a similar remark on the other site (see here).

I realize that Martin Morgan edited his earlier comment. He wanted to refer to the parallel package, not base. So I'll edit the subject here.

You covered parallel in your list, so I'll leave it up to them to reply about it. Maybe there's a use case they have in mind where what you said doesn't hold. But given that install.packages('parallel') fails on R 3.1.2 and 3.2 I have a hard time coming up with one.

[mtmorgan]

My confusion started because parallel was not reported at all (I've revised my original post, empowered by the ability to change the past ;) ). It seems useful to indicate the full user session, including base packages (since both R [thinking of the methods package and Rscript] and the user can do all kinds of funky things, including starting R with no packages other than 'base') and especially base packages like parallel that are not loaded by default in a factory-fresh installation. Also, the user could install an arbitrary version of parallel in their R library directory. Hopefully that utils::sessionInfo() doesn't report versions of these packages is not compelling!

[BrianDiggs]

I hadn't noticed that session_info() did not report the base packages at all; I thought the issue was that the packages were reported without a version number. I do think session_info() should get a include_base argument (default FALSE for backward compatibility, but I could make a completeness argument that TRUE should be the default) which is propagated to the non-exported function package_info which can already handle this.

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/