INNER CODE UNIT · TypeScript
MIDDLE_DOT
vinta/pangu.js · src/shared/index.ts:224
export const MIDDLE_DOT = /(?<![ \u00a0\u00b7\u2022\u2027])[\u00b7\u2022\u2027](?![ \u00a0\u00b7\u2022\u2027])/g;
// A bare unpaired non-void tag amid prose is a tag mention, not markup: it reads as one unit and is spaced from CJK it directly touches
// A trailing self-closing slash is still bare, but void elements render on their own (<br> or <hr>), so they stay markup even unpaired
export const VOID_HTML_TAGS: ReadonlySet<string> = new Set(['area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'track', 'wbr']);
export const BARE_HTML_TAG = /^<([a-zA-Z][a-zA-Z0-9]*)\s*\/?>$/;
export const CLOSING_HTML_TAG = /<\/([a-zA-Z][a-zA-Z0-9]*)/g;
export const CJK_HTML_TAG_MENTION = new RegExp(`([${CJK}])(?=\uE004)`, 'g');
export const HTML_TAG_MENTION_CJK = new RegExp(`(?<=\uE005)([${CJK}])`, 'g');
// A URL reads as one unit: nothing inside it is modified, and it is spaced from CJK on its left. Scheme-anchored only, and CJK characters continue the URL (/wiki/\u4e2d\u6587), so CJK
// prose written tight after a URL stays tight. The body ends at the Private Use Area too, so a URL never swallows a placeholder. See ADR 0026
export const HTTP_URL = /(?<![A-Za-z0-9])https?:\/\/[^\s<>"`\u3000-\u303f\uff00-\uffef\u2018\u2019\u201c\u201d\u2026\ue000-\uf8ff]+/g;
// Trailing half-width punctuation and an unbalanced closing parenthesis belong to the prose, not the URL
export const HTTP_URL_TRAILING_PUNCTUATION = /[.,;:!?'"]+$/;
export const CJK_HTTP_URL = new RegExp(`([${CJK}])(?=\uE00A)`, 'g');