Activate conda environments prior to creating R kernel sessions

[jmcphers]

This change:

• Makes the Conda R setting visible, with an experimental tag.
• Decorates discovered Conda R environments with Conda: env_name in the picker.
• Runs conda activate inside the shell that runs Conda R kernels, using a new kernel supervisor feature. This feature is also intended to support other preflight commands in the future. e.g. module load (Support interpreter environments defined by Environment Modules #8075).

The bulk of this change lives in the kernel supervisor: posit-dev/kallichore#55

Supersedes #10471.

Release Notes

• Experimental support for starting Conda R sessions in Positron via conda activate. Note that not all of Positron's R features have been tested with Conda R. (Support R interpreters installed via conda #4398)

New Features

• N/A

Bug Fixes

• N/A

QA Notes

This change brings in a new supervisor version that has more sophisticated startup handling to support the new features, so it is worth checking that old settings like "Run in Shell" continue to work.

[github-actions]

E2E Tests 🚀
This PR will run tests tagged with: @:critical

^{readme  valid tags}

[wesm]

I tried building this branch and testing it out but couldn't get it working, maybe something wrong with my environment (macOS)

[jmcphers]

@wesm did you run npm install after pulling the branch? (the error looks like a supervisor version mismatch, which would happen if you didn't get the new version of the supervisor along with the rest of the code)

[juliasilge]

I was able to build it successfully, but when I try to start an existing conda R env that I have, I see this:

R 4.4.2 (Conda: my-fantastic-r-env) failed to start up (exit code 1)

The kernel exited before a connection could be established

CommandNotFoundError: Your shell has not been properly configured to use 'conda activate'.
To initialize your shell, run

    $ conda init <SHELL_NAME>

Currently supported shells are:
  - bash
  - fish
  - tcsh
  - xonsh
  - zsh
  - powershell

See 'conda init --help' for more information and options.

IMPORTANT: You may need to close and restart your shell after running 'conda init'.

[juliasilge]

So what I reported in #10788 (comment) was totally a "me" problem; I had cleaned out all the conda activation business out of my dot files at some point. 🙈

Once I did conda init, I could in fact start up my conda R environment and run code, make plots, etc:

I was not able to use install.packages() for a package that contains compiled code, which is the problem reported in #10359. This change does not set things up such that you can use install.packages().

I also want to highlight that it doesn't feel super solid in my experimentation. I experienced several crashes of R along with the R LSP. I do think we should merge this to see how intermittent such problems are (and they may not be related directly) but I wanted to note what I was seeing!

[juliasilge]

Should we make a claim about compilation tools? I still can't use install.packages() with this change and have to install packages via conda.

Also, I do think the suggestion from @jennybc to get more explicit here about "starting the R console" is probably a good idea.