[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 76cd81d60310d65d01f9d7b48a8985d8ab89c8b4)
Co-authored-by: Senthil Kumaran <skumaran@gatech.edu>
(backported to Python 2.7 by Michał Górny)

4 files changed, 55 insertions, 0 deletions

diff --git a/Doc/library/urlparse.rst b/Doc/library/urlparse.rst
index 2f8e4c5a44..0546dcae8c 100644
--- a/Doc/library/urlparse.rst
+++ b/Doc/library/urlparse.rst
@@ -267,6 +267,9 @@ The :mod:`urlparse` module defines the following functions:
    decomposed before parsing, or is not a Unicode string, 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.
+
    .. versionadded:: 2.2
 
    .. versionchanged:: 2.5
@@ -276,6 +279,10 @@ The :mod:`urlparse` module defines the following functions:
       Characters that affect netloc parsing under NFKC normalization will
       now raise :exc:`ValueError`.
 
+   .. versionchanged:: 2.7.18_p9 (Gentoo)
+      ASCII newline and tab characters are stripped from the URL.
+
+.. _WHATWG spec: https://url.spec.whatwg.org/#concept-basic-url-parser
 
 .. function:: urlunsplit(parts)
 
@@ -327,6 +334,10 @@ The :mod:`urlparse` module defines the following functions:
 
 .. 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 urlparse module
       should conform to this. Certain deviations could be observed, which are
@@ -351,6 +362,8 @@ The :mod:`urlparse` module defines the following functions:
    :rfc:`1738` - Uniform Resource Locators (URL)
       This specifies the formal syntax and semantics of absolute URLs.
 
+.. _WHATWG: https://url.spec.whatwg.org/
+
 
 .. _urlparse-result-object:
 
diff --git a/Lib/test/test_urlparse.py b/Lib/test/test_urlparse.py
index 0b2107339a..9912fe2888 100644
--- a/Lib/test/test_urlparse.py
+++ b/Lib/test/test_urlparse.py
@@ -543,6 +543,35 @@ class UrlParseTestCase(unittest.TestCase):
         self.assertEqual(p1.params, 'phone-context=+1-914-555')
 
 
+    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 = urlparse.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 = urlparse.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 non-integer ports."""
         p = urlparse.urlsplit("http://www.example.net:foo")
diff --git a/Lib/urlparse.py b/Lib/urlparse.py
index 6c32727fce..4f24cc12c6 100644
--- a/Lib/urlparse.py
+++ b/Lib/urlparse.py
@@ -62,6 +62,9 @@ scheme_chars = ('abcdefghijklmnopqrstuvwxyz'
                 '0123456789'
                 '+-.')
 
+# Unsafe bytes to be removed per WHATWG spec
+_UNSAFE_URL_BYTES_TO_REMOVE = ['\t', '\r', '\n']
+
 MAX_CACHE_SIZE = 20
 _parse_cache = {}
 
@@ -198,6 +201,10 @@ def urlsplit(url, scheme='', allow_fragments=True):
     if len(_parse_cache) >= MAX_CACHE_SIZE: # avoid runaway growth
         clear_cache()
     netloc = query = fragment = ''
+
+    for b in _UNSAFE_URL_BYTES_TO_REMOVE:
+        url = url.replace(b, "")
+
     i = url.find(':')
     if i > 0:
         if url[:i] == 'http': # optimize the common case
diff --git a/Misc/NEWS.d/next/Security/2021-04-25-07-46-37.bpo-43882.Jpwx85.rst b/Misc/NEWS.d/next/Security/2021-04-25-07-46-37.bpo-43882.Jpwx85.rst
new file mode 100644
index 0000000000..a326d079df
--- /dev/null
+++ b/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.