fix(utils): allow any non-boundary characters in Markdown heading ID (#7604)

packages/docusaurus-utils/src/__tests__/markdownUtils.test.ts

@@ -833,6 +833,53 @@ describe('parseMarkdownHeadingId', () => {
       id: 'id',
     });
   });
+
+  it('does not parse empty id', () => {
+    expect(parseMarkdownHeadingId('## a {#}')).toEqual({
+      text: '## a {#}',
+      id: undefined,
+    });
+  });
+
+  it('can parse id with more characters', () => {
+    expect(parseMarkdownHeadingId('## a {#你好}')).toEqual({
+      text: '## a',
+      id: '你好',
+    });
+
+    expect(parseMarkdownHeadingId('## a {#2022.1.1}')).toEqual({
+      text: '## a',
+      id: '2022.1.1',
+    });
+
+    expect(parseMarkdownHeadingId('## a {#a#b}')).toEqual({
+      text: '## a',
+      id: 'a#b',
+    });
+  });
+
+  // The actual behavior is unspecified, just need to ensure it stays consistent
+  it('handles unmatched boundaries', () => {
+    expect(parseMarkdownHeadingId('## a {# a {#bcd}')).toEqual({
+      text: '## a {# a',
+      id: 'bcd',
+    });
+
+    expect(parseMarkdownHeadingId('## a {#bcd}}')).toEqual({
+      text: '## a {#bcd}}',
+      id: undefined,
+    });
+
+    expect(parseMarkdownHeadingId('## a {#b{cd}')).toEqual({
+      text: '## a',
+      id: 'b{cd',
+    });
+
+    expect(parseMarkdownHeadingId('## a {#b{#b}')).toEqual({
+      text: '## a {#b',
+      id: 'b',
+    });
+  });
 });
 
 describe('writeMarkdownHeadingId', () => {

packages/docusaurus-utils/src/markdownUtils.ts

@@ -14,8 +14,8 @@ import {createSlugger, type Slugger, type SluggerOptions} from './slugger';
 // content. Most parsing is still done in MDX through the mdx-loader.
 
 /**
- * Parses custom ID from a heading. The ID must be composed of letters,
+ * Parses custom ID from a heading. The ID can contain any characters except
- * underscores, and dashes only.
+ * `{#` and `}`.
  *
  * @param heading e.g. `## Some heading {#some-heading}` where the last
  * character must be `}` for the ID to be recognized
@@ -26,9 +26,9 @@ export function parseMarkdownHeadingId(heading: string): {
    */
   text: string;
   /** The heading ID. e.g. `some-heading` */
-  id?: string;
+  id: string | undefined;
 } {
-  const customHeadingIdRegex = /\s*\{#(?<id>[\w-]+)\}$/;
+  const customHeadingIdRegex = /\s*\{#(?<id>(?:.(?!\{#|\}))*.)\}$/;
   const matches = customHeadingIdRegex.exec(heading);
   if (matches) {
     return {

website/_dogfooding/_pages tests/markdownPageTests.md

@@ -162,6 +162,16 @@ function Clock(props) {
 
 ## Custom heading ID {#custom}
 
+### Weird heading {#你好}
+
+### Weird heading {#2022.1.1}
+
+### Weird heading {#a#b}
+
+### Weird heading {#a b}
+
+### Weird heading {#a{b}
+
 ## Pipe
 
 Code tag + double pipe: <code>&#124;&#124;</code>