I've noticed that Clojure multiline docstrings seem to be manually formatted in most cases, including the ones in clojure.core. Example from https://github.com/clojure/clojure/blob/master/src/clj/clojure/core.clj :
(defn flatten
"Takes any nested combination of sequential things (lists, vectors,
etc.) and returns their contents as a single, flat sequence.
(flatten nil) returns an empty sequence."
{:added "1.2"
:static true}
[x]
(filter (complement sequential?)
(rest (tree-seq sequential? seq x))))
This seems odd, as it means that different docstrings will have different line wrap lengths etc. which need to be manually maintained.
Is there a better way to format multiline docstrings?
If you're using Emacs, grab clojure-mode.el
from technomancy's Github, which differs from the one in ELPA (I don't know why, both claim to be version 1.11.5, maybe someone can comment on that?) but includes clojure-fill-docstring
which will format docstrings with nice indentation and linewrapping, bound by default to C-c M-q
.
It will take this:
(defn flatten
"Takes any nested combination of sequential things (lists, vectors, etc.) and returns their contents as a single, flat sequence. (flatten nil) returns an empty sequence."
{:added "1.2"
:static true}
[x]
(filter (complement sequential?)
(rest (tree-seq sequential? seq x))))
and turn it into this:
(defn flatten
"Takes any nested combination of sequential things (lists, vectors,
etc.) and returns their contents as a single, flat sequence.
(flatten nil) returns an empty sequence."
{:added "1.2"
:static true}
[x]
(filter (complement sequential?)
(rest (tree-seq sequential? seq x))))
after you do C-c M-q
with your point inside the docstring.