MD089 - CJK spacing¶
Aliases: cjk-spacing
Disabled by default. This rule is opt-in: enable it explicitly with extend-enable.
What this rule does¶
Reports a Chinese, Japanese or Korean letter that touches a Latin letter or a
digit, and inserts one space between them: 日本語english becomes
日本語 english, 花了5000元 becomes 花了 5000 元, 한글english becomes
한글 english.
Latin text means ASCII letters and digits. Full-width letters and digits
(１２３, Ｔｅｓｔ), letters of other scripts and all punctuation are left
alone: 中文。english and iPhone，好 are correct as written.
The rule reads every line of prose, including headings, list items, table cells and blockquotes. It never changes:
- fenced and indented code blocks, and the inside of inline code, inline math, links, wikilinks and bare URLs; these are spaced from the outside, so a CJK letter touching a code span or a link is separated from it and the span itself is left untouched
- images, HTML tags, HTML comments,
#tagsand footnote markers ([^1]), on either side; a link whose text is one image ([](target)) is a badge and counts as an image - front matter, math blocks, HTML blocks and link reference definitions
- Pandoc and kramdown attribute metadata:
`code`{.class},[text]{.class}, a block IAL line ({:.class}) and the body of a kramdown extension block; a bracketed span carrying attributes is left alone as a whole, so the space before it is not reported either
Emphasis markers are looked through, and the space lands outside them:
**中文**english becomes **中文** english.
Strikethrough is not looked through: ~ is neither an emphasis marker nor a
configured symbol, so a CJK letter touching a strikethrough run
(中文~~english~~) is left alone.
A configured symbol counts as part of the Latin text next to it, so
角度為90°的角 becomes 角度為 90° 的角 and 價格$5 becomes 價格 $5. A symbol
between two CJK words (你好-世界, 注意:这是) is not touched.
A gap whose space would complete a list marker is not reported: 1)中文 is an
enumeration label inside a paragraph, and spacing it would turn the line into
an ordered list item.
Why this matters¶
Chinese, Japanese and Korean copywriting guidelines call for a space between CJK text and Latin words or numbers because the two scripts have no word boundary of their own; the space keeps mixed text readable. Japanese technical writing largely prefers no space, which is why the rule is opt-in.
Examples¶
Incorrect¶
Correct¶
Configuration¶
[MD089]
# The rule is opt-in, so a configuration section alone does not turn it on
enabled = true
# Symbols that lead a Latin run and take a space after a CJK letter: 價格$5 -> 價格 $5
symbols-after-cjk = "-+'\"([¥$"
# Symbols that trail a Latin run and take a space before a CJK letter: 90°的 -> 90° 的
symbols-before-cjk = "-+;:'\"°%$)]"
Each option is one string of characters; whitespace in it is ignored. Set an
option to "" to stop every symbol on that side from attaching.
Automatic fixes¶
The fix inserts one space at each reported position. Existing spaces are never changed; a run of several spaces is MD064's concern.