mui · Janpot · Oct 29, 2025 · Oct 23, 2025 · Oct 23, 2025 · Oct 23, 2025
diff --git a/packages/code-infra/package.json b/packages/code-infra/package.json
@@ -39,6 +39,10 @@
     "./stylelint": {
       "types": "./build/stylelint/index.d.mts",
       "default": "./src/stylelint/index.mjs"
+    },
+    "./brokenLinksChecker": {
+      "types": "./build/brokenLinksChecker/index.d.mts",
+      "default": "./src/brokenLinksChecker/index.mjs"
     }
   },
   "bin": {
@@ -79,6 +83,7 @@
     "babel-plugin-transform-remove-imports": "^1.8.1",
     "chalk": "^5.6.2",
     "clipboardy": "^5.0.0",
+    "content-type": "^1.0.5",
     "env-ci": "^11.2.0",
     "eslint-config-prettier": "^10.1.8",
     "eslint-import-resolver-typescript": "^4.4.4",
@@ -96,6 +101,7 @@
     "globals": "^16.4.0",
     "globby": "^15.0.0",
     "minimatch": "^10.0.3",
+    "node-html-parser": "^7.0.1",
     "open": "^10.2.0",
     "postcss-styled-syntax": "^0.7.1",
     "regexp.escape": "^2.0.1",
@@ -106,14 +112,16 @@
     "yargs": "^18.0.0"
   },
   "peerDependencies": {
+    "@next/eslint-plugin-next": "*",
     "eslint": "^9.0.0",
     "prettier": "^3.5.3",
-    "typescript": "^5.0.0",
-    "@next/eslint-plugin-next": "*"
+    "typescript": "^5.0.0"
   },
   "devDependencies": {
+    "@octokit/types": "^15.0.1",
     "@types/babel__core": "^7.20.5",
     "@types/babel__preset-env": "^7.10.0",
+    "@types/content-type": "^1.1.9",
     "@types/env-ci": "^3.1.4",
     "@types/eslint-plugin-jsx-a11y": "^6.10.1",
     "@types/estree": "^1.0.8",
@@ -123,8 +131,9 @@
     "@typescript-eslint/parser": "^8.46.2",
     "@typescript-eslint/rule-tester": "^8.46.2",
     "eslint": "^9.38.0",
-    "@octokit/types": "^15.0.1",
+    "get-port": "^7.1.0",
     "prettier": "^3.6.2",
+    "serve": "^14.2.5",
     "typescript-eslint": "^8.46.2"
   },
   "files": [

diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/broken-links.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/broken-links.html
@@ -0,0 +1,20 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Page with Broken Links</title>
+  </head>
+  <body>
+    <h1>Page with Broken Links</h1>
+    <p>This page contains links to non-existent pages.</p>
+    <nav>
+      <ul>
+        <li><a href="/">Home (valid)</a></li>
+        <li><a href="/does-not-exist.html">This page does not exist</a></li>
+        <li><a href="/another-missing-page.html">Another missing page</a></li>
+        <li><a href="/valid.html">Valid link</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/broken-targets.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/broken-targets.html
@@ -0,0 +1,22 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Page with Broken Targets</title>
+  </head>
+  <body>
+    <h1>Page with Broken Targets</h1>
+    <p>This page contains links to valid pages but with invalid hash targets.</p>
+    <nav>
+      <ul>
+        <li><a href="/">Home (valid)</a></li>
+        <li><a href="/with-anchors.html#section1">Valid anchor</a></li>
+        <li><a href="/with-anchors.html#nonexistent">Non-existent anchor</a></li>
+        <li><a href="/valid.html#missing-target">Valid page, missing target</a></li>
+        <li><a href="/with-anchors.html#section2">Another valid anchor</a></li>
+        <li><a href="/with-anchors.html#also-missing">Also missing</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/example.md b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/example.md
@@ -0,0 +1,9 @@
+# Example Markdown File
+
+This is a markdown file with an HTML code snippet:
+
+```html
+<a href="/this-should-not-be-checked.html">This link is in a code snippet</a>
+```
+
+This link should not be crawled because this is a markdown file, not HTML.
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/external-links.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/external-links.html
@@ -0,0 +1,21 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Page with External Links</title>
+  </head>
+  <body>
+    <h1>Page with External Links</h1>
+    <p>This page contains external links that should be ignored.</p>
+    <nav>
+      <ul>
+        <li><a href="/">Home (valid internal)</a></li>
+        <li><a href="https://example.com">External link 1</a></li>
+        <li><a href="https://github.com/mui">External link 2</a></li>
+        <li><a href="http://localhost:9999/fake">External localhost</a></li>
+        <li><a href="/valid.html">Valid internal link</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/ignored-page.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/ignored-page.html
@@ -0,0 +1,17 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Ignored Page</title>
+  </head>
+  <body>
+    <h1>Ignored Page</h1>
+    <p>This page should be ignored by the crawler.</p>
+    <nav>
+      <ul>
+        <li><a href="/this-link-should-not-be-checked.html">Should not be checked</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/index.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/index.html
@@ -0,0 +1,26 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Test Site Home</title>
+  </head>
+  <body>
+    <h1>Test Site Home</h1>
+    <nav>
+      <ul>
+        <li><a href="/valid.html">Valid Page</a></li>
+        <li><a href="/with-anchors.html">Page with Anchors</a></li>
+        <li><a href="/broken-links.html">Page with Broken Links</a></li>
+        <li><a href="/broken-targets.html">Page with Broken Targets</a></li>
+        <li><a href="/external-links.html">Page with External Links</a></li>
+        <li><a href="/nested/page.html">Nested Page</a></li>
+        <li><a href="/ignored-page.html">Ignored Page</a></li>
+        <li><a href="/page-with-ignored-content.html">Page with Ignored Content</a></li>
+        <li><a href="/page-with-custom-targets.html">Page with Custom Targets</a></li>
+        <li><a href="/page-with-known-target-links.html">Page with Known Target Links</a></li>
+        <li><a href="/page-with-api-links.html">Page with API Links</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/known-targets.json b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/known-targets.json
@@ -0,0 +1,5 @@
+{
+  "targets": {
+    "/api-page.html": ["#method1", "#method2", "#method3"]
+  }
+}
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/nested/page.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/nested/page.html
@@ -0,0 +1,19 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Nested Page</title>
+  </head>
+  <body>
+    <h1>Nested Page</h1>
+    <p>This is a page in a nested directory.</p>
+    <nav>
+      <ul>
+        <li><a href="/">Home</a></li>
+        <li><a href="/valid.html">Valid page</a></li>
+        <li><a href="/with-anchors.html#section1">Link to anchor</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/orphaned-page.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/orphaned-page.html
@@ -0,0 +1,20 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Orphaned Page</title>
+  </head>
+  <body>
+    <h1>Orphaned Page</h1>
+    <p>This page is not linked from anywhere and can only be discovered via seedUrls.</p>
+    <nav>
+      <ul>
+        <li><a href="/valid.html">Valid Page</a></li>
+        <li><a href="/valid.html/">Valid Page with trailing slash</a></li>
+        <li><a href="/orphaned-broken-link.html">Broken link from orphaned page</a></li>
+        <li><a href="/example.md">Markdown file with HTML snippet</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/page-with-api-links.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/page-with-api-links.html
@@ -0,0 +1,20 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Page with API Links</title>
+  </head>
+  <body>
+    <h1>Page with API Links</h1>
+    <p>This page links to API documentation with known targets.</p>
+    <nav>
+      <ul>
+        <li><a href="/">Home</a></li>
+        <li><a href="/api-page.html#method1">API Method 1 (known)</a></li>
+        <li><a href="/api-page.html#method2">API Method 2 (known)</a></li>
+        <li><a href="/api-page.html#unknown-method">Unknown API method</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/.../code-infra/src/brokenLinksChecker/__fixtures__/static-site/page-with-custom-targets.html b/.../code-infra/src/brokenLinksChecker/__fixtures__/static-site/page-with-custom-targets.html
@@ -0,0 +1,24 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Page with Custom Targets</title>
+  </head>
+  <body>
+    <h1>Page with Custom Targets</h1>
+    <nav>
+      <ul>
+        <li><a href="/">Home</a></li>
+        <li><a href="#custom-id">Link to custom ID</a></li>
+        <li><a href="#__should-be-ignored">Link to ignored ID</a></li>
+      </ul>
+    </nav>
+    <section id="custom-id">
+      <h2>Custom ID Section</h2>
+    </section>
+    <section id="__should-be-ignored">
+      <h2>This target should be ignored</h2>
+    </section>
+  </body>
+</html>
diff --git a/...code-infra/src/brokenLinksChecker/__fixtures__/static-site/page-with-ignored-content.html b/...code-infra/src/brokenLinksChecker/__fixtures__/static-site/page-with-ignored-content.html
@@ -0,0 +1,28 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Page with Ignored Content</title>
+  </head>
+  <body>
+    <h1>Page with Ignored Content</h1>
+    <nav>
+      <ul>
+        <li><a href="/">Home</a></li>
+      </ul>
+    </nav>
+    <aside class="sidebar">
+      <h2>Sidebar (should be ignored)</h2>
+      <ul>
+        <li><a href="/sidebar-broken-link.html">Broken link in sidebar</a></li>
+      </ul>
+    </aside>
+    <main>
+      <p>Main content</p>
+      <p>
+        <a href="/broken-link-with-sidebar-class.html" class="sidebar">Link with sidebar class directly</a>
+      </p>
+    </main>
+  </body>
+</html>
diff --git a/...e-infra/src/brokenLinksChecker/__fixtures__/static-site/page-with-known-target-links.html b/...e-infra/src/brokenLinksChecker/__fixtures__/static-site/page-with-known-target-links.html
@@ -0,0 +1,19 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Page with Known Target Links</title>
+  </head>
+  <body>
+    <h1>Page with Known Target Links</h1>
+    <p>This page links to external pages with known targets.</p>
+    <nav>
+      <ul>
+        <li><a href="/">Home</a></li>
+        <li><a href="/external-page.html#valid-target">Valid external target</a></li>
+        <li><a href="/external-page.html#invalid-target">Invalid external target</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/valid.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/valid.html
@@ -0,0 +1,20 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Valid Page</title>
+  </head>
+  <body>
+    <h1>Valid Page</h1>
+    <p>This page has only valid internal links.</p>
+    <nav>
+      <ul>
+        <li><a href="/">Home</a></li>
+        <li><a href="/with-anchors.html#section1">Link to section 1</a></li>
+        <li><a href="/with-anchors.html#section2">Link to section 2</a></li>
+        <li><a href="/nested/page.html">Nested page</a></li>
+      </ul>
+    </nav>
+  </body>
+</html>
diff --git a/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/with-anchors.html b/packages/code-infra/src/brokenLinksChecker/__fixtures__/static-site/with-anchors.html
@@ -0,0 +1,31 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Page with Anchors</title>
+  </head>
+  <body>
+    <h1>Page with Anchors</h1>
+    <nav>
+      <ul>
+        <li><a href="/">Home</a></li>
+        <li><a href="#section1">Jump to section 1</a></li>
+        <li><a href="#section2">Jump to section 2</a></li>
+        <li><a href="#section3">Jump to section 3</a></li>
+      </ul>
+    </nav>
+    <section id="section1">
+      <h2>Section 1</h2>
+      <p>Content for section 1</p>
+    </section>
+    <section id="section2">
+      <h2>Section 2</h2>
+      <p>Content for section 2</p>
+    </section>
+    <section id="section3">
+      <h2>Section 3</h2>
+      <p>Content for section 3</p>
+    </section>
+  </body>
+</html>