Update ess layer readme by common configuration options

[rommeswi]

The commit adds two configuration options to the readme.

The first is a way to stop ESS from asking what the R working directory should be.
The second is a way to stop ESS from indenting standard R comments starting with '#' by 40 spaces.

I think a lot of ESS users will be interested in knowing these configuration options.
It took me a while to find them on the web, therefore I think it would be useful to include them in the layer readme.
For the 40 space automatic indentation, I even thought it was a bug.

[ryanprior]

This looks helpful, thank you!

It's interesting that CircleCI is failing on "cat: /tmp/changed_files: No such file or directory" - I'm guessing that's from the end of this script: https://github.com/syl20bnr/spacemacs/blob/develop/.circleci/select_pr_changed#L23

As an experiment @rommeswi, will you make a small update to an .el file and see if that allows the documentation test to run?
@JAremko do you have any idea why it's failing like that?

[JAremko]

@ryanprior looks like one of the involved services glitched.

[rommeswi]

Circleci passed immediately with the change in the .el file (before it failed on the first attempt). However, it was also makerd as passed right before I added the other commit. So I am not sure if something has been fixed, it was just a glitch, or indeed editing an .el file helps.

[downward-funarg]

It should be noted that setting ess-ask-for-ess-directory true causes the REPL to open in the project home directory, not the file directory. See emacs-ess/ESS#461

[rommeswi]

I think that "feature" no longer exists: Ess Manual
I also just confirmed that on emacs itself. I started the process in a subdirectory of a git folder and R used setwd on the subdirectory, not the project directory.

(Also, I think you meant "false" instead of "true", or? If the value is set to true then it will ask instead of choosing the home directory.)

[downward-funarg]

Sorry, yes I meant "false". But I can't reproduce your finding:

(setq-default ess-ask-for-ess-directory nil)
(make-directory "~/test-R/test" t)
(magit-init "~/test-R")
(find-file "~/test-R/test/repl-test.R")
(save-buffer)
(spacemacs/ess-start-repl)

opens in ~/test-R.

[rommeswi]

Strange. I just tried again. I am in a file inside a git repository that also shows up in my home buffer. My modeline shows me that it indeed recognizes that it is inside a git repo. No matter whether I then run M-x R or M-X spacemacs/ess-start-repl, I always get setwd to the folder where the R file is located, not the git repo.

The thread that you linked is 6 years old and it would be undocumented behavior (see the link I posted above). Are you sure you are on a current version of ESS and emacs? I could also imagine that it matters in which folder you execute the lisp code you posted above.

In any case, if you have any additions to make to the readme, please go ahead and create a separate pull request.

[downward-funarg]

This is indeed strange, I just tried with a fresh clone of spacemacs on a different emacs version and OS (30.0.50, Ubuntu 20.04; my initial system was emacs-mac-port on macOS Sonoma) and got the same behavior that I reported.

But I guess this is not really spacemacs' problem, except insofar as this behavior is prevalent enough to mention in the readme - would be useful if others can give their experience to decide about that.

[downward-funarg]

@rommeswi Actually, opening in the project directory is expected behavior and is documented. From the documentation for ess-directory:

If ess-startup-directory-function is nil, then the current project as returned by project-current is used.

And it's in the ESS manual:

ESS[R]: Add support for R projects and start R by default in the project folder.

I wonder why your setup doesn't do this! Anyway, the docstring also suggests setting (setq ess-startup-directory 'default-directory) to disable this behavior, which works for me.

[rommeswi]

@downward-funarg please go ahead and create a pull request with an updated readme that properly documents behavior in your view. I don't think this is a productive discussion. The behavior on my system that I tried to enter into the readme is consistent with what is linked in the manual link I provided above.

[downward-funarg]

I'm commenting on your pull request because your contribution would add an option which might have detrimental or unexpected effects. Feel free to ignore my comments, I suppose, but this information is relevant when considering if this change should be merged.

I understand that your system behaves as you say it does, but so does mine, and if you'll read the manual link which I also provided, you will see that the behavior I describe is expected for the current ESS version.

[rommeswi]

If you can comment, I'm sure you can also update the readme. Go ahead. I'm closing this.