Skip to content

Highlighting, indentation, and subprocess support for the Clojure language

Notifications You must be signed in to change notification settings

jeffvalk/clojure-mode

 
 

Repository files navigation

circleci MELPA MELPA Stable NonGNU ELPA Discord License GPL 3

Clojure Mode

clojure-mode is an Emacs major mode that provides font-lock (syntax highlighting), indentation, navigation and refactoring support for the Clojure(Script) programming language.


This documentation tracks the master branch of clojure-mode. Some of the features and settings discussed here might not be available in older releases (including the current stable release). Please, consult the relevant git tag (e.g. 5.19.0) if you need documentation for a specific clojure-mode release.

Installation

Available on the major package.el community maintained repos - MELPA Stable and MELPA repos.

MELPA Stable is the recommended repo as it has the latest stable version. MELPA has a development snapshot for users who don't mind (infrequent) breakage but don't want to run from a git checkout.

You can install clojure-mode using the following command:

M-x package-install [RET] clojure-mode [RET]

or if you'd rather keep it in your dotfiles:

(unless (package-installed-p 'clojure-mode)
  (package-install 'clojure-mode))

If the installation doesn't work try refreshing the package list:

M-x package-refresh-contents

Bundled major modes

The clojure-mode package actually bundles together several major modes:

  • clojure-mode is a major mode for editing Clojure code
  • clojurescript-mode is a major mode for editing ClojureScript code
  • clojurec-mode is a major mode for editing .cljc source files

All the major modes derive from clojure-mode and provide more or less the same functionality. Differences can be found mostly in the font-locking - e.g. ClojureScript has some built-in constructs that are not present in Clojure.

The proper major mode is selected automatically based on the extension of the file you're editing.

Having separate major modes gives you the flexibility to attach different hooks to them and to alter their behavior individually (e.g. add extra font-locking just to clojurescript-mode) .

Note that all modes derive from clojure-mode, so things you add to clojure-mode-hook and clojure-mode-map will affect all the derived modes as well.

Configuration

In the spirit of Emacs, pretty much everything you can think of in clojure-mode is configurable.

To see a list of available configuration options do M-x customize-group RET clojure.

Indentation options

The default indentation rules in clojure-mode are derived from the community Clojure Style Guide. Please, refer to the guide for the general Clojure indentation rules.

If you'd like to use the alternative "fixed/tonsky" indentation style you should update your configuration accordingly:

(setq clojure-indent-style 'always-indent
      clojure-indent-keyword-style 'always-indent
      clojure-enable-indent-specs nil)

Read on for more details on the available indentation-related configuration options.

Indentation of docstrings

By default multi-line docstrings are indented with 2 spaces, as this is a somewhat common standard in the Clojure community. You can however adjust this by modifying clojure-docstring-fill-prefix-width. Set it to 0 if you don't want multi-line docstrings to be indented at all (which is pretty common in most lisps).

Indentation of function forms

The indentation of function forms is configured by the variable clojure-indent-style. It takes three possible values:

  • always-align (the default)
(some-function
 10
 1
 2)
(some-function 10
               1
               2)
  • always-indent
(some-function
  10
  1
  2)
(some-function 10
  1
  2)
  • align-arguments
(some-function
  10
  1
  2)
(some-function 10
               1
               2)

Note: Prior to clojure-mode 5.10, the configuration options for clojure-indent-style used to be keywords, but now they are symbols. Keywords will still be supported at least until clojure-mode 6.

Indentation of keywords

Similarly we have the clojure-indent-keyword-style, which works in the following way:

  • always-align (default) - All args are vertically aligned with the first arg in case (A), and vertically aligned with the function name in case (B).
(:require [foo.bar]
          [bar.baz])
(:require
 [foo.bar]
 [bar.baz])
  • always-indent - All args are indented like a macro body.
(:require [foo.bar]
   [bar.baz])
(:x
   location
   0)
  • align-arguments - Case (A) is indented like always-align, and case (B) is indented like a macro body.
(:require [foo.bar]
          [bar.baz])
(:x
   location
   0)

Indentation of macro forms

The indentation of special forms and macros with bodies is controlled via put-clojure-indent, define-clojure-indent and clojure-backtracking-indent. Nearly all special forms and built-in macros with bodies have special indentation settings in clojure-mode. You can add/alter the indentation settings in your personal config. Let's assume you want to indent ->> and -> like this:

(->> something
  ala
  bala
  portokala)

You can do so by putting the following in your config:

(put-clojure-indent '-> 1)
(put-clojure-indent '->> 1)

This means that the body of the ->/->> is after the first argument.

A more compact way to do the same thing is:

(define-clojure-indent
  (-> 1)
  (->> 1))

To indent something like a definition (defn) you can do something like:

(put-clojure-indent '>defn :defn)

You can also specify different indentation settings for symbols prefixed with some ns (or ns alias):

(put-clojure-indent 'do 0)
(put-clojure-indent 'my-ns/do 1)

The bodies of certain more complicated macros and special forms (e.g. letfn, deftype, extend-protocol, etc) are indented using a contextual backtracking indentation method, require more sophisticated indent specifications. Here are a few examples:

(define-clojure-indent
  (implement '(1 (1)))
  (letfn     '(1 ((:defn)) nil))
  (proxy     '(2 nil nil (1)))
  (reify     '(:defn (1)))
  (deftype   '(2 nil nil (1)))
  (defrecord '(2 nil nil (1)))
  (specify   '(1 (1)))
  (specify   '(1 (1))))

These follow the same rules as the :style/indent metadata specified by cider-nrepl. For instructions on how to write these specifications, see this document. The only difference is that you're allowed to use lists instead of vectors.

The indentation of special arguments is controlled by clojure-special-arg-indent-factor, which by default indents special arguments a further lisp-body-indent when compared to ordinary arguments.

An example of the default formatting is:

(defrecord MyRecord
    [my-field])

Note that defrecord has two special arguments, followed by the form's body - namely the record's name and its fields vector.

Setting clojure-special-arg-indent-factor to 1, results in:

(defrecord MyRecord
  [my-field])

You can completely disable the effect of indentation specs like this:

(setq clojure-enable-indent-specs nil)

Indentation of Comments

clojure-mode differentiates between comments like ;, ;;, etc. By default clojure-mode treats ; as inline comments and always indents those. You can change this behaviour like this:

(add-hook 'clojure-mode-hook (lambda () (setq-local comment-column 0)))

You might also want to change comment-add to 0 in that way, so that Emacs comment functions (e.g. comment-region) would use ; by default instead of ;;.

Note: Check out this section of the Clojure style guide to understand better the semantics of the different comment levels and why clojure-mode treats them differently by default.

Vertical alignment

You can vertically align sexps with C-c SPC. For instance, typing this combo on the following form:

(def my-map
  {:a-key 1
   :other-key 2})

Leads to the following:

(def my-map
  {:a-key     1
   :other-key 2})

This can also be done automatically (as part of indentation) by turning on clojure-align-forms-automatically. This way it will happen whenever you select some code and hit TAB.

Font-locking

clojure-mode features static font-locking (syntax highlighting) that you can extend yourself if needed. As typical for Emacs, it's based on regular expressions. You can find the default font-locking rules in clojure-font-lock-keywords. Here's how you can add font-locking for built-in Clojure functions and vars:

(defvar clojure-built-in-vars
  '(;; clojure.core
    "accessor" "aclone"
    "agent" "agent-errors" "aget" "alength" "alias"
    "all-ns" "alter" "alter-meta!" "alter-var-root" "amap"
    ;; omitted for brevity
    ))

(defvar clojure-built-in-dynamic-vars
  '(;; clojure.test
    "*initial-report-counters*" "*load-tests*" "*report-counters*"
    "*stack-trace-depth*" "*test-out*" "*testing-contexts*" "*testing-vars*"
    ;; clojure.xml
    "*current*" "*sb*" "*stack*" "*state*"
    ))

(font-lock-add-keywords 'clojure-mode
                        `((,(concat "(\\(?:\.*/\\)?"
                                    (regexp-opt clojure-built-in-vars t)
                                    "\\>")
                           1 font-lock-builtin-face)))

(font-lock-add-keywords 'clojure-mode
                        `((,(concat "\\<"
                                    (regexp-opt clojure-built-in-dynamic-vars t)
                                    "\\>")
                           0 font-lock-builtin-face)))

Note: The package clojure-mode-extra-font-locking provides such additional font-locking for Clojure built-ins.

As you might imagine one problem with this font-locking approach is that because it's based on regular expressions you'll get some false positives here and there (there's no namespace information, and no way for clojure-mode to know what var a symbol resolves to). That's why clojure-mode's font-locking defaults are conservative and minimalistic.

Precise font-locking requires additional data that can obtained from a running REPL (that's how CIDER's dynamic font-locking works) or from static code analysis.

When it comes to non built-in definitions, clojure-mode needs to be manually instructed how to handle the docstrings and highlighting. Here's an example:

(put '>defn 'clojure-doc-string-elt 2)

(font-lock-add-keywords 'clojure-mode
                        `((,(concat "(\\(?:" clojure--sym-regexp "/\\)?"
                                    "\\(>defn\\)\\>")
                           1 font-lock-keyword-face)))

Note: The clojure-doc-string-elt attribute is processed by the function clojure-font-lock-syntactic-face-function.

Refactoring support

The available refactorings were originally created and maintained by the clj-refactor.el team. The ones implemented in Elisp only are gradually migrated to clojure-mode.

Threading macros related features

clojure-thread: Thread another form into the surrounding thread. Both ->> and -> variants are supported.

clojure-unwind: Unwind a threaded expression. Supports both ->> and ->.

clojure-thread-first-all: Introduce the thread first macro (->) and rewrite the entire form. With a prefix argument do not thread the last form.

clojure-thread-last-all: Introduce the thread last macro and rewrite the entire form. With a prefix argument do not thread the last form.

clojure-unwind-all: Fully unwind a threaded expression removing the threading macro.

Cycling things

clojure-cycle-privacy: Cycle privacy of defs or defns. Use metadata explicitly with setting clojure-use-metadata-for-privacy to t for defns too.

clojure-cycle-not: Add or remove a not form around the current form.

clojure-cycle-when: Find the closest when or when-not up the syntax tree and toggle it.

clojure-cycle-if: Find the closest if or if-not up the syntax tree and toggle it. Also transpose the else and then branches, keeping the semantics the same as before.

Convert collection

Convert any given collection at point to list, quoted list, map, vector or set.

Let expression

clojure-introduce-let: Introduce a new let form. Put the current form into its binding form with a name provided by the user as a bound name. If called with a numeric prefix put the let form Nth level up in the form hierarchy.

clojure-move-to-let: Move the current form to the closest let's binding form. Replace all occurrences of the form in the body of the let.

clojure-let-forward-slurp-sexp: Slurp the next form after the let into the let. Replace all occurrences of the bound forms in the form added to the let form. If called with a prefix argument slurp the next n forms.

clojure-let-backward-slurp-sexp: Slurp the form before the let into the let. Replace all occurrences of the bound forms in the form added to the let form. If called with a prefix argument slurp the previous n forms.

paredit-convolute-sexp is advised to replace occurrences of bound forms with their bound names when convolute is used on a let form.

Rename ns alias

clojure-rename-ns-alias: Rename an alias inside a namespace declaration, and all of its usages in the buffer

If there is an active selected region, only rename usages of aliases within the region, without affecting the namespace declaration.

Add arity to a function

clojure-add-arity: Add a new arity to an existing single-arity or multi-arity function.

Related packages

  • clojure-mode-extra-font-locking provides additional font-locking for built-in methods and macros. The font-locking is pretty imprecise, because it doesn't take namespaces into account and it won't font-lock a function at all possible positions in a sexp, but if you don't mind its imperfections you can easily enable it:
(require 'clojure-mode-extra-font-locking)

The code in clojure-mode-font-locking used to be bundled with clojure-mode before version 3.0.

You can also use the code in this package as a basis for extending the font-locking further (e.g. functions/macros from more namespaces). Generally you should avoid adding special font-locking for things that don't have fairly unique names, as this will result in plenty of incorrect font-locking. CIDER users should avoid this package, as CIDER does its own dynamic font-locking, which is namespace-aware and doesn't produce almost any false positives.

  • clj-refactor provides refactoring support.

  • Enabling CamelCase support for editing commands(like forward-word, backward-word, etc) in clojure-mode is quite useful since we often have to deal with Java class and method names. The built-in Emacs minor mode subword-mode provides such functionality:

(add-hook 'clojure-mode-hook #'subword-mode)
  • The use of paredit when editing Clojure (or any other Lisp) code is highly recommended. It helps ensure the structure of your forms is not compromised and offers a number of operations that work on code structure at a higher level than just characters and words. To enable it for Clojure buffers:
(add-hook 'clojure-mode-hook #'paredit-mode)
  • smartparens is an excellent (newer) alternative to paredit. Many Clojure hackers have adopted it recently and you might want to give it a try as well. To enable smartparens use the following code:
(add-hook 'clojure-mode-hook #'smartparens-strict-mode)
  • RainbowDelimiters is a minor mode which highlights parentheses, brackets, and braces according to their depth. Each successive level is highlighted in a different color. This makes it easy to spot matching delimiters, orient yourself in the code, and tell which statements are at a given depth. Assuming you've already installed RainbowDelimiters you can enable it like this:
(add-hook 'clojure-mode-hook #'rainbow-delimiters-mode)
  • aggressive-indent-mode automatically adjust the indentation of your code, while you're writing it. Using it together with clojure-mode is highly recommended. Provided you've already installed aggressive-indent-mode you can enable it like this:
(add-hook 'clojure-mode-hook #'aggressive-indent-mode)

REPL Interaction

One of the fundamental aspects of Lisps in general, and Clojure in particular, is the notion of interactive programming - building your programs by continuously changing the state of the running Lisp program (as opposed to doing something more traditional like making a change and re-running the program afterwards to see the changes in action). To get the most of clojure-mode you'll have to combine it with some tool which will allow you to interact with your Clojure program (a.k.a. process/REPL).

A number of options exist for connecting to a running Clojure process and evaluating code interactively.

Basic REPL

inf-clojure provides basic interaction with a Clojure REPL process. It's very similar in nature and supported functionality to inferior-lisp-mode for Common Lisp.

CIDER

CIDER is a powerful Clojure interactive development environment, similar to SLIME for Common Lisp.

If you're into Clojure and Emacs you should definitely check it out.

Tutorials

Tutorials, targeting Emacs beginners, are available at clojure-doc.org and Clojure for the Brave and the True. Keep in mind, however, that they might be out-of-date.

Caveats

clojure-mode is a capable tool, but it's certainly not perfect. This section lists a couple of general design problems/limitations that might affect your experience negatively.

General Issues

clojure-mode derives a lot of functionality directly from lisp-mode (an Emacs major mode for Common Lisp), which simplified the initial implementation, but also made it harder to implement certain functionality. Down the road it'd be nice to fully decouple clojure-mode from lisp-mode.

See this ticket for a bit more details.

Indentation Performance

clojure-mode's indentation engine is a bit slow. You can speed things up significantly by disabling clojure-use-backtracking-indent, but this will break the indentation of complex forms like deftype, defprotocol, reify, letfn, etc.

We should look into ways to optimize the performance of the backtracking indentation logic. See this ticket for more details.

Font-locking Implementation

As mentioned above, the font-locking is implemented in terms of regular expressions which makes it both slow and inaccurate.

Changelog

An extensive changelog is available here.

License

Copyright © 2007-2024 Jeffrey Chu, Lennart Staflin, Phil Hagelberg, Bozhidar Batsov, Artur Malabarba, Magnar Sveen and contributors.

Distributed under the GNU General Public License; type C-h C-c to view it.

About

Highlighting, indentation, and subprocess support for the Clojure language

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Emacs Lisp 97.6%
  • Clojure 2.3%
  • Makefile 0.1%