Infrastructure: link-checker.js: Check if any of the known pageIds …

…include the hash being checked for (pull #2812) Modifies link checker so that when GitHub dynamically prepends "user-content-" to fragment IDs that are in user-generated content, the link test will pass.
w3c · Oct 2, 2023 · f97050b · f97050b
1 parent 2b619c1
commit f97050b
Show file tree

Hide file tree

Showing 2 changed files with 35 additions and 3 deletions.
diff --git a/.link-checker.js b/.link-checker.js
@@ -11,9 +11,16 @@ module.exports = {
       '.carousel-image a',
     ],
   },
+  hashCheckHandlers: [
+    {
+      name: 'github',
+      pattern: /^https:\/\/github\.com\/.*/,
+      matchHash: (ids, hash) =>
+        ids.includes(hash) || ids.includes(`user-content-${hash}`),
+    },
+  ],
   ignoreHashesOnExternalPagesMatchingRegex: [
     // Some hash links are resolved with JS and are therefore difficult to check algorithmically
-    /^https:\/\/github\.com\/.*\/wiki\//,
     /^https:\/\/html\.spec\.whatwg\.org\/multipage\//,
   ],
 };
diff --git a/scripts/link-checker.js b/scripts/link-checker.js
@@ -33,6 +33,23 @@ async function checkLinks() {
     return getLineNumber;
   };
 
+  const checkPathForHash = (hrefOrSrc, ids = [], hash) => {
+    // On some websites, the ids may not exactly match the hash included
+    // in the link.
+    // For e.g. GitHub will prepend client facing ids with their own
+    // calculated value. A heading in a README for example could be
+    // 'Foo bar', navigated to with https://github.com/foo/bar#foo-bar,
+    // but GitHub calculates the actual markup id included in the document
+    // as being 'user-content-foo-bar' for its own page processing purposes.
+    //
+    // See https://github.com/w3c/aria-practices/issues/2809
+    const handler = options.hashCheckHandlers.find(({ pattern }) =>
+      pattern.test(hrefOrSrc)
+    );
+    if (handler) return handler.matchHash(ids, hash);
+    else return ids.includes(hash);
+  };
+
   const countConsoleErrors = () => {
     let errorCount = 0;
 
@@ -227,7 +244,11 @@ async function checkLinks() {
 
         let matchesHash = true;
         if (hash) {
-          matchesHash = !!matchingPage?.ids.includes(hash);
+          matchesHash = !!checkPathForHash(
+            pathMinusHash,
+            matchingPage?.ids,
+            hash
+          );
         }
 
         const isLinkBroken = !(
@@ -274,7 +295,11 @@ async function checkLinks() {
             hrefOrSrc.match(pattern)
           );
 
-        if (!isHashCheckingDisabled && hash && !pageData.ids.includes(hash)) {
+        if (
+          !isHashCheckingDisabled &&
+          hash &&
+          !checkPathForHash(hrefOrSrc, pageData.ids, hash)
+        ) {
           consoleError(
             `Found broken external link on ${htmlPath}:${lineNumber}:${columnNumber}, ` +
               'hash not found on page'