Do not continue crawling on the other origin when the origin-crossing…

… relation results in an HTTP redirect #196 (comment)
Munter · Nov 21, 2021 · 240f584 · 240f584
1 parent a6ef642
commit 240f584
Show file tree

Hide file tree

Showing 3 changed files with 61 additions and 2 deletions.
diff --git a/lib/index.js b/lib/index.js
@@ -647,7 +647,11 @@ async function hyperlink(
         } else if (
           ['HtmlAnchor', 'SvgAnchor', 'HtmlIFrame'].includes(relation.type)
         ) {
-          if (!relation.crossorigin && recursive) {
+          if (
+            !relation.crossorigin &&
+            !relation.from.crossedOrigins &&
+            recursive
+          ) {
             follow = true;
           } else if (relation.from !== relation.to) {
             // If we are handling local file-urls, follow but mark as end-of-line in processing
@@ -689,7 +693,8 @@ async function hyperlink(
 
         if (follow) {
           // Save information about cross origin navigations for later
-          relation.to.crossedOrigins = relation.crossorigin;
+          relation.to.crossedOrigins =
+            relation.crossorigin || relation.from.crossedOrigins;
 
           if (assetTypesWithoutRelations.includes(relation.to.type)) {
             // If we are handling local file-urls, follow but mark as end-of-line in processing

diff --git a/test/index.js b/test/index.js
@@ -1048,6 +1048,59 @@ describe('hyperlink', function () {
         });
       });
     });
+
+    // Regression test for https://github.com/Munter/hyperlink/issues/196#issuecomment-974726167
+    it('should not follow relations after a redirect upon already crossing origins', async function () {
+      httpception([
+        {
+          request: 'GET https://docs.ovh.com/gb/en/dedicated/firewall-network',
+          response: {
+            statusCode: 301,
+            headers: {
+              'Content-Type': 'text/html',
+              Location:
+                'https://docs.ovh.com/gb/en/dedicated/firewall-network/',
+            },
+            body: '<html><head><title>301 Moved Permanently</title></head></html>',
+          },
+        },
+        {
+          request: 'GET https://docs.ovh.com/gb/en/dedicated/firewall-network/',
+          response: {
+            statusCode: 200,
+            headers: { 'Content-Type': 'text/html' },
+            body: `<html><head></head><body><a id="objective" href="/otherpage.html"></body></html>`,
+          },
+        },
+      ]);
+
+      const t = new TapRender();
+      sinon.spy(t, 'push');
+      await hyperlink(
+        {
+          recursive: true,
+          root: pathModule.resolve(
+            __dirname,
+            '..',
+            'testdata',
+            'externalLinkRedirect'
+          ),
+          inputUrls: ['index.html'],
+        },
+        t
+      );
+      expect(t.close(), 'to satisfy', { fail: 1 });
+      expect(t.push, 'to have a call satisfying', () => {
+        t.push(null, {
+          operator: 'fragment-check',
+          name: 'fragment-check testdata/externalLinkRedirect/index.html --> https://docs.ovh.com/gb/en/dedicated/firewall-network#objective',
+          expected: 'id="objective"',
+          at: 'testdata/externalLinkRedirect/index.html:1:10 <a href="https://docs.ovh.com/gb/en/dedicated/firewall-network#objective">...</a>',
+          ok: true,
+          actual: 'id="objective"',
+        });
+      });
+    });
   });
 
   describe('with a relation that points at an asset that returns 404', function () {

diff --git a/testdata/externalLinkRedirect/index.html b/testdata/externalLinkRedirect/index.html
@@ -0,0 +1 @@
+<a href="https://docs.ovh.com/gb/en/dedicated/firewall-network#objective">hey</a>
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1 @@
		<a href="https://docs.ovh.com/gb/en/dedicated/firewall-network#objective">hey</a>