Setting up Scheme Development in Emacs

Esperanto ■ English
Last updated: March 16, 2022

All the good ideas never lie under one hat.
―Dale Turner


Table of contents


In this article, I’ll discuss the easiest approach that I took to setup up a Scheme development environment in Emacs. Take note, that this is not the only approach available—some did it in arguably better ways. In this article, I’ll try to explain the shortest route that I took.

Editing Scheme code with Emacs has traditionally been done by rudimentary modes that lacked flexibility. They were able to evaluate current definitions, last definitions, and entire buffers, for the most part. Unfortunately, that didn’t suffice with the way Scheme dealt with things. A more intelligent way of handling code, was needed.

Fortunately, there is Geiser. There are other major modes that try to do what Geiser does, but I became most comfortable with what Geiser offered. Some similar libraries can co-exist with Geiser, too. I tried those, but it became too complex, for me. I wound up just using Geiser. Also, as a semi-related note, I’m using Emacs to edit Scheme code because I don’t know of any other editor that does it so well.


With ELPA, installing Geiser is a breeze. Simply execute the following

M-x package-install RET geiser RET

And in a few seconds, you’ll have Geiser installed in your Emacs profile. Next, put in the actual code that invokes and configures Geiser:

(require 'geiser)

(setq geiser-active-implementations '(mit))

(defun geiser-save ()

The first expression loads Geiser, itself. The second one specifies that it won’t prompt you for other implementations if it finds them. The last one is optional—it enables you to execute

M-x geiser-save RET

in the REPL buffer to force saving of the history to the disk file, which is found in ~/, by default. It is useful if you want to save your REPL session, immediately. Nothing is more horrifying than losing THAT expression. For all the Emacs code above, to take effect, you can evaluate them now using members of the EVAL troupe—eval-defun, eval-last-sexp, eval-region—or, you can still opt to respawn a new Emacs process.


To reap what you sowed, create or open a .scm file, with at least a proper module declaration. Then hit:

M-x run-geiser RET

And, boomshakalaka! A new (Emacs) window opens, containing the * MIT REPL * buffer. Whatever you can do with the REPL invoked with vanilla command-line mit-scheme, you can also do with this, and more. This major mode is actually Comint mode, under the hood, with hooks to a Scheme process. For those of you who are unfamiliar with Comint mode, it is the same mode that handles `M-x shell RET`.

So, what can you do with it? While editing .scm file, here are some of the usual shortcuts that I use. The full list of keys are available here. Take note, that the description of the keys that I use below, are for myself initially, to help me understand what they do. They may, or may not diverge from the official description, listed on the aforementioned link.

Scheme buffer

KeyWhat it does
C-c C-zSwitch to the REPL buffer
C-c C-aEvaluate current buffer, then switch to the REPL buffer
C-M-xEvaluate current expression
C-x C-eEvaluate last expression
C-c C-rEvaluate region
C-c C-\Insert a lambda (λ) symbol

REPL buffer

KeyWhat it does
C-c C-zSwitch to the Scheme buffer
M-pSwitch to the previous history item
M-nSwitch to the next history item
C-c M-pJump to previous prompt
C-c M-nJump to next prompt
C-c C-qQuit the REPL

Closing remarks

I have intentionally skipped many topics from the official document because it makes it unattractive to people who are averse to reading long blocks of text. Ironically, this article may even qualify as one. The methods described above are by in no way representative of how the general community does them. See ya!