fix: improve link rewrite (#2213)

fixes #2212
adobe · Jul 14, 2023 · e4a4ce7 · e4a4ce7
1 parent 1b39601
commit e4a4ce7
Show file tree

Hide file tree

Showing 3 changed files with 64 additions and 36 deletions.
diff --git a/src/server/HelixImportServer.js b/src/server/HelixImportServer.js
@@ -134,9 +134,12 @@ export class HelixImportServer extends BaseServer {
       }
     }
 
-    let buffer;
+    let buffer = await ret.buffer();
+    if (contentType.includes('html') || contentType.includes('text')) {
+      buffer = utils.rewriteUrl(buffer, host);
+    }
+
     if (!isBodyReq && this._project.cacheDirectory) {
-      buffer = await ret.buffer();
       await utils.writeToCache(
         url,
         this._project.cacheDirectory,
@@ -149,40 +152,12 @@ export class HelixImportServer extends BaseServer {
       );
     }
 
-    if (contentType.includes('html') || contentType.includes('text')) {
-      // make urls relative
-      let text;
-      if (buffer) {
-        text = buffer.toString();
-      } else {
-        text = await ret.text();
-      }
-      let replacer = new RegExp(`${host}/`, 'gm');
-      text = text.replace(replacer, '/');
-      replacer = new RegExp(host, 'gm');
-      text = text.replace(replacer, '/');
-      res
-        .status(ret.status)
-        .set(respHeaders)
-        .cookie('hlx-proxyhost', host)
-        .send(text);
-      ret.body.pipe(res);
-      return;
-    }
-
-    if (buffer) {
-      res
-        .status(ret.status)
-        .set(respHeaders)
-        .cookie('hlx-proxyhost', host)
-        .send(buffer);
-      ret.body.pipe(res);
-    } else {
-      res
-        .status(ret.status)
-        .set(respHeaders);
-      ret.body.pipe(res);
-    }
+    res
+      .status(ret.status)
+      .set(respHeaders)
+      .cookie('hlx-proxyhost', host)
+      .send(buffer);
+    ret.body.pipe(res);
   }
 
   /**

diff --git a/src/server/utils.js b/src/server/utils.js
@@ -438,6 +438,20 @@ window.LiveReloadOptions = {
       });
     });
   },
+
+  /**
+   * Rewrites all absolute urls to the proxy host with relative ones.
+   * @param {Buffer} html
+   * @param {string} host
+   * @returns {Buffer}
+   */
+  rewriteUrl(html, host) {
+    const hostPattern = host.replaceAll('.', '\\.');
+    let text = html.toString('utf-8');
+    const re = new RegExp(`(src|href)\\s*=\\s*(["'])${hostPattern}(/.*?)?(['"])`, 'gm');
+    text = text.replaceAll(re, (match, arg, q1, value, q2) => (`${arg}=${q1}${value || '/'}${q2}`));
+    return Buffer.from(text, 'utf-8');
+  },
 };
 
 export default Object.freeze(utils);
diff --git a/test/server-utils.test.js b/test/server-utils.test.js
@@ -280,4 +280,43 @@ describe('Utils Test', () => {
       });
     });
   });
+
+  describe('URL Rewrite', () => {
+    function test(html, expected) {
+      const actual = utils.rewriteUrl(Buffer.from(html), 'https://www.example.com').toString('utf-8');
+      assert.strictEqual(actual, expected);
+    }
+
+    it('replaces host in src', () => {
+      test('<img src="https://www.example.com/foo.png">', '<img src="/foo.png">');
+    });
+
+    it('replaces host in src (single quotes)', () => {
+      test("<img src = 'https://www.example.com/foo.png' width='200'>", "<img src='/foo.png' width='200'>");
+    });
+
+    it('replaces host in href', () => {
+      test('<a href="https://www.example.com">', '<a href="/">');
+    });
+
+    it('replaces host in href with slash', () => {
+      test('<a href="https://www.example.com/">', '<a href="/">');
+    });
+
+    it('replaces host in href', () => {
+      test('<a href="https://www.example.com">', '<a href="/">');
+    });
+
+    it('does proper matching', () => {
+      test('<a href="https://www.example.comm">', '<a href="https://www.example.comm">');
+    });
+
+    it('does not replace host in query params', () => {
+      test('<a href="https://facebook.com?u=https://www.example.com">', '<a href="https://facebook.com?u=https://www.example.com">');
+    });
+
+    it('does not replace the host in text', () => {
+      test('<a href="https://www.example.com">https://www.example.com</a>', '<a href="/">https://www.example.com</a>');
+    });
+  });
 });