[3.9] bpo-43882 - urllib.parse should sanitize urls containing ASCII …

…newline and tabs. (GH-25595) (GH-25725)

* bpo-43882 - urllib.parse should sanitize urls containing ASCII newline and tabs. (GH-25595)

Co-authored-by: Gregory P. Smith <greg@krypto.org>
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
(cherry picked from commit 76cd81d)
Co-authored-by: Senthil Kumaran <skumaran@gatech.edu>

Doc/library/urllib.parse.rst

@@ -311,6 +311,9 @@ or on combining URL components into a URL string.
    ``#``, ``@``, or ``:`` will raise a :exc:`ValueError`. If the URL is
    decomposed before parsing, no error will be raised.
 
+   Following the `WHATWG spec`_ that updates RFC 3986, ASCII newline
+   ``\n``, ``\r`` and tab ``\t`` characters are stripped from the URL.
+
    .. versionchanged:: 3.6
       Out-of-range port numbers now raise :exc:`ValueError`, instead of
       returning :const:`None`.
@@ -319,6 +322,10 @@ or on combining URL components into a URL string.
       Characters that affect netloc parsing under NFKC normalization will
       now raise :exc:`ValueError`.
 
+   .. versionchanged:: 3.9.5
+      ASCII newline and tab characters are stripped from the URL.
+
+.. _WHATWG spec: https://url.spec.whatwg.org/#concept-basic-url-parser
 
 .. function:: urlunsplit(parts)
 
@@ -673,6 +680,10 @@ task isn't already covered by the URL parsing functions above.
 
 .. seealso::
 
+   `WHATWG`_ -  URL Living standard
+      Working Group for the URL Standard that defines URLs, domains, IP addresses, the
+      application/x-www-form-urlencoded format, and their API.
+
    :rfc:`3986` - Uniform Resource Identifiers
       This is the current standard (STD66). Any changes to urllib.parse module
       should conform to this. Certain deviations could be observed, which are
@@ -696,3 +707,5 @@ task isn't already covered by the URL parsing functions above.
 
    :rfc:`1738` - Uniform Resource Locators (URL)
       This specifies the formal syntax and semantics of absolute URLs.
+
+.. _WHATWG: https://url.spec.whatwg.org/

Lib/test/test_urlparse.py

@@ -612,6 +612,35 @@ def test_urlsplit_attributes(self):
         with self.assertRaisesRegex(ValueError, "out of range"):
             p.port
 
+    def test_urlsplit_remove_unsafe_bytes(self):
+        # Remove ASCII tabs and newlines from input
+        url = "http://www.python.org/java\nscript:\talert('msg\r\n')/#frag"
+        p = urllib.parse.urlsplit(url)
+        self.assertEqual(p.scheme, "http")
+        self.assertEqual(p.netloc, "www.python.org")
+        self.assertEqual(p.path, "/javascript:alert('msg')/")
+        self.assertEqual(p.query, "")
+        self.assertEqual(p.fragment, "frag")
+        self.assertEqual(p.username, None)
+        self.assertEqual(p.password, None)
+        self.assertEqual(p.hostname, "www.python.org")
+        self.assertEqual(p.port, None)
+        self.assertEqual(p.geturl(), "http://www.python.org/javascript:alert('msg')/#frag")
+
+        # Remove ASCII tabs and newlines from input as bytes.
+        url = b"http://www.python.org/java\nscript:\talert('msg\r\n')/#frag"
+        p = urllib.parse.urlsplit(url)
+        self.assertEqual(p.scheme, b"http")
+        self.assertEqual(p.netloc, b"www.python.org")
+        self.assertEqual(p.path, b"/javascript:alert('msg')/")
+        self.assertEqual(p.query, b"")
+        self.assertEqual(p.fragment, b"frag")
+        self.assertEqual(p.username, None)
+        self.assertEqual(p.password, None)
+        self.assertEqual(p.hostname, b"www.python.org")
+        self.assertEqual(p.port, None)
+        self.assertEqual(p.geturl(), b"http://www.python.org/javascript:alert('msg')/#frag")
+
     def test_attributes_bad_port(self):
         """Check handling of invalid ports."""
         for bytes in (False, True):

Lib/urllib/parse.py

@@ -78,6 +78,9 @@
                 '0123456789'
                 '+-.')
 
+# Unsafe bytes to be removed per WHATWG spec
+_UNSAFE_URL_BYTES_TO_REMOVE = ['\t', '\r', '\n']
+
 # XXX: Consider replacing with functools.lru_cache
 MAX_CACHE_SIZE = 20
 _parse_cache = {}
@@ -469,6 +472,9 @@ def urlsplit(url, scheme='', allow_fragments=True):
         else:
             scheme, url = url[:i].lower(), url[i+1:]
 
+    for b in _UNSAFE_URL_BYTES_TO_REMOVE:
+        url = url.replace(b, "")
+
     if url[:2] == '//':
         netloc, url = _splitnetloc(url, 2)
         if (('[' in netloc and ']' not in netloc) or

Misc/NEWS.d/next/Security/2021-04-25-07-46-37.bpo-43882.Jpwx85.rst

@@ -0,0 +1,6 @@
+The presence of newline or tab characters in parts of a URL could allow
+some forms of attacks.
+
+Following the controlling specification for URLs defined by WHATWG
+:func:`urllib.parse` now removes ASCII newlines and tabs from URLs,
+preventing such attacks.