Update URL sanitizer to allow more protocols (#38531)

Co-authored-by: XhmikosR <[email protected]>

1 files changed, 68 insertions, 10 deletions

diff --git a/js/tests/unit/util/sanitizer.spec.js b/js/tests/unit/util/sanitizer.spec.js
index 55e9b6336..2b21ef2e1 100644
--- a/js/tests/unit/util/sanitizer.spec.js
+++ b/js/tests/unit/util/sanitizer.spec.js
@@ -10,17 +10,75 @@ describe('Sanitizer', () => {
       expect(result).toEqual(empty)
     })
 
-    it('should sanitize template by removing tags with XSS', () => {
-      const template = [
-        '<div>',
-        '  <a href="javascript:alert(7)">Click me</a>',
-        '  <span>Some content</span>',
-        '</div>'
-      ].join('')
-
-      const result = sanitizeHtml(template, DefaultAllowlist, null)
+    it('should retain tags with valid URLs', () => {
+      const validUrls = [
+        '',
+        'http://abc',
+        'HTTP://abc',
+        'https://abc',
+        'HTTPS://abc',
+        'ftp://abc',
+        'FTP://abc',
+        'mailto:[email protected]',
+        'MAILTO:[email protected]',
+        'tel:123-123-1234',
+        'TEL:123-123-1234',
+        'sip:[email protected]',
+        'SIP:[email protected]',
+        '#anchor',
+        '/page1.md',
+        'http://JavaScript/my.js',
+        'data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAABAAAAAQCAYAAAAf8/', // Truncated.
+        'data:video/webm;base64,iVBORw0KGgoAAAANSUhEUgAAABAAAAAQCAYAAAAf8/',
+        'data:audio/opus;base64,iVBORw0KGgoAAAANSUhEUgAAABAAAAAQCAYAAAAf8/',
+        'unknown-scheme:abc'
+      ]
+
+      for (const url of validUrls) {
+        const template = [
+          '<div>',
+          `  <a href="${url}">Click me</a>`,
+          '  <span>Some content</span>',
+          '</div>'
+        ].join('')
+
+        const result = sanitizeHtml(template, DefaultAllowlist, null)
+
+        expect(result).toContain(`href="${url}"`)
+      }
+    })
 
-      expect(result).not.toContain('href="javascript:alert(7)')
+    it('should sanitize template by removing tags with XSS', () => {
+      const invalidUrls = [
+        // eslint-disable-next-line no-script-url
+        'javascript:alert(7)',
+        // eslint-disable-next-line no-script-url
+        'javascript:evil()',
+        // eslint-disable-next-line no-script-url
+        'JavaScript:abc',
+        ' javascript:abc',
+        ' \n Java\n Script:abc',
+        '&#106;&#97;&#118;&#97;&#115;&#99;&#114;&#105;&#112;&#116;&#58;',
+        '&#106&#97;&#118;&#97;&#115;&#99;&#114;&#105;&#112;&#116;&#58;',
+        '&#106 &#97;&#118;&#97;&#115;&#99;&#114;&#105;&#112;&#116;&#58;',
+        '&#0000106&#0000097&#0000118&#0000097&#0000115&#0000099&#0000114&#0000105&#0000112&#0000116&#0000058',
+        '&#x6A&#x61&#x76&#x61&#x73&#x63&#x72&#x69&#x70&#x74&#x3A;',
+        'jav&#x09;ascript:alert();',
+        'jav\u0000ascript:alert();'
+      ]
+
+      for (const url of invalidUrls) {
+        const template = [
+          '<div>',
+          `  <a href="${url}">Click me</a>`,
+          '  <span>Some content</span>',
+          '</div>'
+        ].join('')
+
+        const result = sanitizeHtml(template, DefaultAllowlist, null)
+
+        expect(result).not.toContain(`href="${url}"`)
+      }
     })
 
     it('should sanitize template and work with multiple regex', () => {