[Dayshine]

For one, in my experience, CRAN only stores binaries of the most recent release of each package. This means that either you have to accept you can never rebuild a docker image, or you have to make sure that you are always able to recompile all of your R packages from source.

This means you don't just have to pin your R package versions, you have to pin all the build dependencies.

And you have to have a different image for different sets of R packages because they might have different build dependencies.

[claytonjy]

Reproducibility is of course a valid concern, but I never had issues with that thanks to Rocker and MRAN.

I do sympathize with long from-source build times; as a Linux user I don't think I had binaries available until I stopped using R, so I've spent days, perhaps weeks, waiting for dplyr to install over the course of my R usage.

> different image for different sets of R packages because they might have different build dependencies

Is this not true of all software in all languages?