bpo-30500: urllib: Simplify splithost by calling into urlparse. (#1849) (#2294)

vstinner · web-flow · commit d4324baca4c0 · 2017-06-20T16:20:36.000+02:00
The current regex based splitting produces a wrong result. For example:: http://abc#@def Web browsers parse that URL as ``http://abc/#@def``, that is, the host is ``abc``, the path is ``/``, and the fragment is ``#@def``. (cherry picked from commit 90e01e5)
diff --git a/Lib/test/test_urllib.py b/Lib/test/test_urllib.py
@@ -879,6 +879,26 @@ def test_splithost(self):
         self.assertEqual(splithost('/foo/bar/baz.html'),
                          (None, '/foo/bar/baz.html'))
 
+        # bpo-30500: # starts a fragment.
+        self.assertEqual(splithost('//127.0.0.1#@host.com'),
+                         ('127.0.0.1', '/#@host.com'))
+        self.assertEqual(splithost('//127.0.0.1#@host.com:80'),
+                         ('127.0.0.1', '/#@host.com:80'))
+        self.assertEqual(splithost('//127.0.0.1:80#@host.com'),
+                         ('127.0.0.1:80', '/#@host.com'))
+
+        # Empty host is returned as empty string.
+        self.assertEqual(splithost("///file"),
+                         ('', '/file'))
+
+        # Trailing semicolon, question mark and hash symbol are kept.
+        self.assertEqual(splithost("//example.net/file;"),
+                         ('example.net', '/file;'))
+        self.assertEqual(splithost("//example.net/file?"),
+                         ('example.net', '/file?'))
+        self.assertEqual(splithost("//example.net/file#"),
+                         ('example.net', '/file#'))
+
     def test_splituser(self):
         splituser = urllib.splituser
         self.assertEqual(splituser('User:Pass@www.python.org:080'),
diff --git a/Lib/urllib.py b/Lib/urllib.py
@@ -1093,8 +1093,7 @@ def splithost(url):
     """splithost('//host[:port]/path') --> 'host[:port]', '/path'."""
     global _hostprog
     if _hostprog is None:
-        import re
-        _hostprog = re.compile('^//([^/?]*)(.*)$')
+        _hostprog = re.compile('//([^/#?]*)(.*)', re.DOTALL)
 
     match = _hostprog.match(url)
     if match:
diff --git a/Misc/ACKS b/Misc/ACKS
@@ -993,6 +993,7 @@ Chad Netzer
 Max Neunhöffer
 George Neville-Neil
 Hieu Nguyen
+Nam Nguyen
 Johannes Nicolai
 Samuel Nicolary
 Jonathan Niehof
diff --git a/Misc/NEWS b/Misc/NEWS
@@ -52,6 +52,11 @@ Extension Modules
 Library
 -------
 
+- [Security] bpo-30500: Fix urllib.splithost() to correctly parse
+  fragments. For example, ``splithost('//127.0.0.1#@evil.com/')`` now
+  correctly returns the ``127.0.0.1`` host, instead of treating ``@evil.com``
+  as the host in an authentification (``login@host``).
+
 - [Security] bpo-29591: Update expat copy from 2.1.1 to 2.2.0 to get fixes
   of CVE-2016-0718 and CVE-2016-4472. See
   https://sourceforge.net/p/expat/bugs/537/ for more information.
