@cffnpwr/textlint-rule-no-dash

A textlint rule that detects the use of dashes (— ― –) in body text, headings, and table cells.

日本語のREADMEはこちら

It encourages rewriting appositive or supplementary insertions with dashes (A——挿入——B) using parentheses（）, and rephrasings (A——B) either by splitting them into two sentences with a full stop or by joining them with a comma.

<!-- NG -->
この機能——つまり自動保存——は便利です。
結論は明快—シンプルが最善です。

<!-- OK -->
この機能（自動保存）は便利です。
結論は明快です。シンプルが最善です。

By default, the en dash (– U+2013) is detected only when both sides are Japanese characters (hiragana, katakana, kanji, ー, 々). English compounds such as Curry–Howard and range notations such as 1–3月 are not detected.

Detection is performed on the body text with markup such as emphasis and links removed. Even when a dash sits on a markup boundary, as in **強調**–続き, the detection condition is judged by the surrounding body characters (「調」 and 「続」 in this example). Inline code, code blocks, HTML tags, images, and autolink (<https://…>) URLs do not have their internal dashes detected. These are replaced with dummy characters of the same length for the judgment, so in the Japanese-character check for adjacent dashes they are treated as non-Japanese characters (for example, the en dash in カリー–codeです。 is not detected because inline code immediately follows it). Links in the [表示テキスト](URL) form target the display text for detection, not the URL. Consecutive dashes (such as ——) are reported as a single error.

How to Install

npm

npm install --save-dev @cffnpwr/textlint-rule-no-dash

or

npx jsr add -D @cffnpwr/textlint-rule-no-dash

yarn

yarn add --dev @cffnpwr/textlint-rule-no-dash

or

yarn dlx jsr add -D @cffnpwr/textlint-rule-no-dash

pnpm

pnpm add --save-dev @cffnpwr/textlint-rule-no-dash

or

pnpm dlx jsr add -D @cffnpwr/textlint-rule-no-dash

Bun

bun add --dev @cffnpwr/textlint-rule-no-dash

or

bunx jsr add -D @cffnpwr/textlint-rule-no-dash

Deno

deno add --dev npm:@cffnpwr/textlint-rule-no-dash

or

deno add --dev jsr:@cffnpwr/textlint-rule-no-dash

How to Use

Add it to .textlintrc.json.

{
  "rules": {
    "@cffnpwr/textlint-rule-no-dash": true
  }
}

Options

type DashContext = "always" | "japanese-both" | "japanese-either";

interface Options {
  allows?: string[];
  dashes?: Partial<Record<"emDash" | "horizontalBar" | "enDash", DashContext>>;
  skipBlockQuote?: boolean;
}

| Option | Default | Description | | --- | --- | --- | | allows | [] | A list of strings or patterns to exclude from detection. In the regexp-string-matcher format, regular expressions can also be specified via "/pattern/". Detections overlapping a matched range are skipped | | dashes | see below | A map of dashes to detect and their detection conditions. When specified, it overrides the entire default (it is not merged per key) | | skipBlockQuote | true | Excludes content under quotes (BlockQuote) from inspection |

The keys of dashes and their corresponding characters are as follows.

| Key | Character | Default detection condition | | --- | --- | --- | | emDash | — U+2014 | always | | horizontalBar | ― U+2015 | always | | enDash | – U+2013 | japanese-both |

The meanings of the detection conditions (DashContext) are as follows.

| Value | Meaning | | --- | --- | | always | Detects unconditionally | | japanese-both | Detects only when both the preceding and following characters are Japanese | | japanese-either | Detects only when at least one of the preceding and following characters is Japanese |

Japanese characters refers to hiragana, katakana, kanji, the prolonged sound mark 「ー」, and 「々」. Specifying an unknown option key or an invalid value results in an error.

{
  "rules": {
    "@cffnpwr/textlint-rule-no-dash": {
      "allows": ["カリー–ハワード"],
      "dashes": {
        "emDash": "always",
        "horizontalBar": "always"
      }
    }
  }
}

License

MIT