From 90d86e1cbf41238431cc92c27091599ce8411dab Mon Sep 17 00:00:00 2001
From: Om Shivaprasad Wanere <omwanere593@gmail.com>
Date: Tue, 13 Jan 2026 21:06:32 +0530
Subject: [PATCH] Fix canonical audit to classify invalid URLs correctly

---
 core/audits/seo/canonical.js           | 30 +++++++++++++++++++++-----
 core/test/audits/seo/canonical-test.js | 27 +++++++++++++++++++++++
 2 files changed, 52 insertions(+), 5 deletions(-)

diff --git a/core/audits/seo/canonical.js b/core/audits/seo/canonical.js
index 1a1868188e4b..a8f9617cc6f9 100644
--- a/core/audits/seo/canonical.js
+++ b/core/audits/seo/canonical.js
@@ -91,11 +91,31 @@ class Canonical extends Audit {
         if (!link.hrefRaw) continue;
 
         // Links that had an hrefRaw but didn't have a valid href were invalid, flag them
-        if (!link.href) invalidCanonicalLink = link;
-        // Links that had a valid href but didn't have a valid hrefRaw must have been relatively resolved, flag them
-        else if (!UrlUtils.isValid(link.hrefRaw)) relativeCanonicallink = link;
-        // Otherwise, it was a valid canonical URL
-        else uniqueCanonicalURLs.add(link.href);
+        // if (!link.href) invalidCanonicalLink = link;
+        // // Links that had a valid href but didn't have a valid hrefRaw must have been relatively resolved, flag them
+        // else if (!UrlUtils.isValid(link.hrefRaw)) relativeCanonicallink = link;
+        // // Otherwise, it was a valid canonical URL
+        // else uniqueCanonicalURLs.add(link.href);
+
+        if (!link.href) {
+          invalidCanonicalLink = link;
+        } else {
+          let parsed;
+          try {
+            parsed = new URL(link.hrefRaw);
+          } catch {
+            invalidCanonicalLink = link;
+            continue;
+          }
+
+          if (!parsed.origin || parsed.origin === 'null') {
+            relativeCanonicallink = link;
+          } else {
+            uniqueCanonicalURLs.add(link.href);
+          }
+        }
+
+
       } else if (link.rel === 'alternate') {
         if (link.href && link.hreflang) hreflangURLs.add(link.href);
       }
diff --git a/core/test/audits/seo/canonical-test.js b/core/test/audits/seo/canonical-test.js
index bf335317d7dc..51db4b277276 100644
--- a/core/test/audits/seo/canonical-test.js
+++ b/core/test/audits/seo/canonical-test.js
@@ -103,6 +103,33 @@ describe('SEO: Document has valid canonical link', () => {
     });
   });
 
+  it('fails when canonical url is invalid', async () => {
+  const mainDocumentUrl = 'https://example.com/';
+  const mainResource = {url: mainDocumentUrl};
+  const devtoolsLog = networkRecordsToDevtoolsLog([mainResource]);
+
+  const artifacts = {
+    DevtoolsLog: devtoolsLog,
+    URL: {mainDocumentUrl},
+    LinkElements: [
+      link({
+        rel: 'canonical',
+        source: 'headers',
+        href: null,
+        hrefRaw: 'example.com',
+      }),
+    ],
+  };
+
+  const context = createContext();
+  const result = await Canonical.audit(artifacts, context);
+
+  expect(result.score).toBe(0);
+  expect(result.explanation)
+    .toContain('Invalid URL');
+});
+
+
   it('fails when canonical points to a different hreflang', () => {
     const mainDocumentUrl = 'https://example.com/';
     const mainResource = {url: mainDocumentUrl};