`R_HOME` is more for introspection than configuration

I'm pretty intrigued by this project! I saw it in use during Hadley's recent data science hangout.

So I came to look at the README so I can try it out.

I've done a lot of work on R runtime discovery in Positron and one thing I read caught my eye:

https://github.com/posit-dev/mcp-repl#r-interpreter-which-r-installation-is-used

To force a specific R installation, set R_HOME in the environment that launches mcp-repl.

We've occasionally seen folks trying to use R_HOME like this in Positron's issues and it generally leads to frustration. While trying to understand those situations, I've had to read up on what R_HOME is designed to do and I think this is basically fighting against the intent of R_HOME:

https://github.com/posit-dev/positron/issues/12278#issuecomment-4001245549

TL;DR

My impression is that R_HOME is intended to be a env var you can consult once you've launched and R version to ask it "where do you live?" It's not something one sets and then expects that to drive the usage of a specific R installation.

I just thought this might be worth some 🤔

Review the README section "R interpreter: which R installation is used" and the linked discussion about R_HOME. Determine whether the guidance should be revised to reflect the reported behavior, then verify that the README clearly explains how an R installation is selected.