[4.2.x] Fixed CVE-2024-41991 -- Prevented potential ReDoS in django.utils.html.urlize() and AdminURLFieldWidget.

felixxm · sarahboyce · sarahboyce · commit efea1ef7e219 · 2024-07-31T16:12:23.000+02:00
Thanks Seokchan Yoon for the report.

Co-authored-by: Sarah Boyce &lt;42296566+sarahboyce@users.noreply.github.com&gt;
diff --git a/django/contrib/admin/widgets.py b/django/contrib/admin/widgets.py
@@ -383,7 +383,7 @@ def get_context(self, name, value, attrs):
         context["current_label"] = _("Currently:")
         context["change_label"] = _("Change:")
         context["widget"]["href"] = (
-            smart_urlquote(context["widget"]["value"]) if value else ""
+            smart_urlquote(context["widget"]["value"]) if url_valid else ""
         )
         context["url_valid"] = url_valid
         return context
diff --git a/django/utils/html.py b/django/utils/html.py
@@ -13,6 +13,8 @@
 from django.utils.safestring import SafeData, SafeString, mark_safe
 from django.utils.text import normalize_newlines
 
+MAX_URL_LENGTH = 2048
+
 
 @keep_lazy(SafeString)
 def escape(text):
@@ -300,9 +302,9 @@ def handle_word(
             # Make URL we want to point to.
             url = None
             nofollow_attr = ' rel="nofollow"' if nofollow else ""
-            if self.simple_url_re.match(middle):
+            if len(middle) <= MAX_URL_LENGTH and self.simple_url_re.match(middle):
                 url = smart_urlquote(html.unescape(middle))
-            elif self.simple_url_2_re.match(middle):
+            elif len(middle) <= MAX_URL_LENGTH and self.simple_url_2_re.match(middle):
                 url = smart_urlquote("http://%s" % html.unescape(middle))
             elif ":" not in middle and self.is_email_simple(middle):
                 local, domain = middle.rsplit("@", 1)
@@ -417,6 +419,10 @@ def is_email_simple(value):
         except ValueError:
             # value contains more than one @.
             return False
+        # Max length for domain name labels is 63 characters per RFC 1034.
+        # Helps to avoid ReDoS vectors in the domain part.
+        if len(p2) > 63:
+            return False
         # Dot must be in p2 (e.g. example.com)
         if "." not in p2 or p2.startswith("."):
             return False
diff --git a/docs/releases/4.2.15.txt b/docs/releases/4.2.15.txt
@@ -23,6 +23,13 @@ CVE-2024-41990: Potential denial-of-service vulnerability in ``django.utils.html
 denial-of-service attack via very large inputs with a specific sequence of
 characters.
 
+CVE-2024-41991: Potential denial-of-service vulnerability in ``django.utils.html.urlize()`` and ``AdminURLFieldWidget``
+=======================================================================================================================
+
+:tfilter:`urlize`, :tfilter:`urlizetrunc`, and ``AdminURLFieldWidget`` were
+subject to a potential denial-of-service attack via certain inputs with a very
+large number of Unicode characters.
+
 Bugfixes
 ========
 
diff --git a/tests/admin_widgets/tests.py b/tests/admin_widgets/tests.py
@@ -461,7 +461,12 @@ def test_localization(self):
 class AdminURLWidgetTest(SimpleTestCase):
     def test_get_context_validates_url(self):
         w = widgets.AdminURLFieldWidget()
-        for invalid in ["", "/not/a/full/url/", 'javascript:alert("Danger XSS!")']:
+        for invalid in [
+            "",
+            "/not/a/full/url/",
+            'javascript:alert("Danger XSS!")',
+            "http://" + "&#54620;.&#44544;." * 1_000_000 + "com",
+        ]:
             with self.subTest(url=invalid):
                 self.assertFalse(w.get_context("name", invalid, {})["url_valid"])
         self.assertTrue(w.get_context("name", "http://example.com", {})["url_valid"])
diff --git a/tests/utils_tests/test_html.py b/tests/utils_tests/test_html.py
@@ -328,6 +328,15 @@ def test_urlize(self):
                 'Search for <a href="http://google.com/?q=">google.com/?q=</a>!',
             ),
             ("foo@example.com", '<a href="mailto:foo@example.com">foo@example.com</a>'),
+            (
+                "test@" + "&#54620;.&#44544;." * 15 + "aaa",
+                '<a href="mailto:test@'
+                + "xn--6q8b.xn--bj0b." * 15
+                + 'aaa">'
+                + "test@"
+                + "&#54620;.&#44544;." * 15
+                + "aaa</a>",
+            ),
         )
         for value, output in tests:
             with self.subTest(value=value):
@@ -336,6 +345,10 @@ def test_urlize(self):
     def test_urlize_unchanged_inputs(self):
         tests = (
             ("a" + "@a" * 50000) + "a",  # simple_email_re catastrophic test
+            # Unicode domain catastrophic tests.
+            "a@" + "&#54620;.&#44544;." * 1_000_000 + "a",
+            "http://" + "&#54620;.&#44544;." * 1_000_000 + "com",
+            "www." + "&#54620;.&#44544;." * 1_000_000 + "com",
             ("a" + "." * 1000000) + "a",  # trailing_punctuation catastrophic test
             "foo@",
             "@foo.com",

Original file line number	Diff line number	Diff line change
`@@ -383,7 +383,7 @@ def get_context(self, name, value, attrs):`
`383`	`383`	`context["current_label"] = _("Currently:")`
`384`	`384`	`context["change_label"] = _("Change:")`
`385`	`385`	`context["widget"]["href"] = (`
`386`		`- smart_urlquote(context["widget"]["value"]) if value else ""`
	`386`	`+ smart_urlquote(context["widget"]["value"]) if url_valid else ""`
`387`	`387`	`)`
`388`	`388`	`context["url_valid"] = url_valid`
`389`	`389`	`return context`