copy frontmatter parsing code from prettier base code

Author: taniki

src/utils/frontmatter/constants.js

@@ -0,0 +1,3 @@
+/** @type {unique symbol} */
+export const FRONT_MATTER_MARK = Symbol.for("PRETTIER_IS_FRONT_MATTER");
+export const FRONT_MATTER_VISITOR_KEYS = [];

src/utils/frontmatter/parse.js

@@ -0,0 +1,121 @@
+import replaceNonLineBreaksWithSpace from './replace-non-line-breaks-with-space.js';
+import { FRONT_MATTER_MARK } from './constants.js';
+
+const DELIMITER_LENGTH = 3;
+
+/**
+@typedef {{
+  index: number,
+  // 1-based line number
+  line: number,
+  // 0-based column number
+  column: number,
+}} Position
+@typedef {{
+  language: string,
+  explicitLanguage: string | null,
+  value: string,
+  startDelimiter: string,
+  endDelimiter: string,
+  raw: string,
+  start: Position,
+  end: Position,
+  [FRONT_MATTER_MARK]: true,
+}} FrontMatter
+*/
+
+/**
+@param {string} text
+@returns {FrontMatter | undefined}
+*/
+function getFrontMatter(text) {
+  const startDelimiter = text.slice(0, DELIMITER_LENGTH);
+
+  if (startDelimiter !== '---' && startDelimiter !== '+++') {
+    return;
+  }
+
+  const firstLineBreakIndex = text.indexOf('\n', DELIMITER_LENGTH);
+  if (firstLineBreakIndex === -1) {
+    return;
+  }
+
+  const explicitLanguage = text
+    .slice(DELIMITER_LENGTH, firstLineBreakIndex)
+    .trim();
+
+  let endDelimiterIndex = text.indexOf(
+    `\n${startDelimiter}`,
+    firstLineBreakIndex,
+  );
+
+  let language = explicitLanguage;
+  if (!language) {
+    language = startDelimiter === '+++' ? 'toml' : 'yaml';
+  }
+
+  if (
+    endDelimiterIndex === -1 &&
+    startDelimiter === '---' &&
+    language === 'yaml'
+  ) {
+    // In some markdown processors such as pandoc,
+    // "..." can be used as the end delimiter for YAML front-matter.
+    endDelimiterIndex = text.indexOf('\n...', firstLineBreakIndex);
+  }
+
+  if (endDelimiterIndex === -1) {
+    return;
+  }
+
+  const frontMatterEndIndex = endDelimiterIndex + 1 + DELIMITER_LENGTH;
+
+  const nextCharacter = text.charAt(frontMatterEndIndex + 1);
+  if (!/\s?/.test(nextCharacter)) {
+    return;
+  }
+
+  const raw = text.slice(0, frontMatterEndIndex);
+  /** @type {string[]} */
+  let lines;
+
+  return {
+    language,
+    explicitLanguage: explicitLanguage || null,
+    value: text.slice(firstLineBreakIndex + 1, endDelimiterIndex),
+    startDelimiter,
+    endDelimiter: raw.slice(-DELIMITER_LENGTH),
+    raw,
+    start: { line: 1, column: 0, index: 0 },
+    end: {
+      index: raw.length,
+      get line() {
+        lines ??= raw.split('\n');
+        return lines.length;
+      },
+      get column() {
+        lines ??= raw.split('\n');
+        return lines.at(-1).length;
+      },
+    },
+    [FRONT_MATTER_MARK]: true,
+  };
+}
+
+function parse(text) {
+  const frontMatter = getFrontMatter(text);
+
+  if (!frontMatter) {
+    return { content: text };
+  }
+
+  return {
+    frontMatter,
+    get content() {
+      const { raw } = frontMatter;
+      return replaceNonLineBreaksWithSpace(raw) + text.slice(raw.length);
+    },
+  };
+}
+
+export default parse;

src/utils/frontmatter/print.js

@@ -0,0 +1,5 @@
+function printFrontMatter({ node }) {
+  return node.raw;
+}
+
+export default printFrontMatter;

src/utils/frontmatter/replace-non-line-breaks-with-space.js

@@ -0,0 +1,19 @@
+// import * as assert from "#universal/assert";
+
+/**
+Replaces all characters in the input string except line breaks `\n` with a space.
+
+@param {string} string - The input string to process.
+@returns {string}
+*/
+function replaceNonLineBreaksWithSpace(string) {
+  const replaced = string.replaceAll(/[^\n]/g, ' ');
+
+  if (process.env.NODE_ENV !== 'production') {
+    // assert.equal(replaced.length, string.length);
+  }
+
+  return replaced;
+}
+
+export default replaceNonLineBreaksWithSpace;