Fixed #15617 - CSRF referer checking too strict

Thanks to adam for the report.

git-svn-id: http://code.djangoproject.com/svn/django/trunk@15840 bcc190cf-cafb-0310-a4f2-bffc1f526a37

django/middleware/csrf.py

@@ -13,6 +13,7 @@ from django.conf import settings
 from django.core.urlresolvers import get_callable
 from django.utils.cache import patch_vary_headers
 from django.utils.hashcompat import md5_constructor
+from django.utils.http import same_origin
 from django.utils.log import getLogger
 from django.utils.safestring import mark_safe
 from django.utils.crypto import constant_time_compare
@@ -161,10 +162,9 @@ class CsrfViewMiddleware(object):
                     )
                     return self._reject(request, REASON_NO_REFERER)
 
-                # The following check ensures that the referer is HTTPS,
-                # the domains match and the ports match - the same origin policy.
+                # Note that request.get_host() includes the port
                 good_referer = 'https://%s/' % request.get_host()
-                if not referer.startswith(good_referer):
+                if not same_origin(referer, good_referer):
                     reason = REASON_BAD_REFERER % (referer, good_referer)
                     logger.warning('Forbidden (%s): %s' % (reason, request.path),
                         extra={

django/utils/http.py

@@ -3,6 +3,7 @@ import datetime
 import re
 import sys
 import urllib
+import urlparse
 from email.Utils import formatdate
 
 from django.utils.encoding import smart_str, force_unicode
@@ -186,3 +187,20 @@ def quote_etag(etag):
     """
     return '"%s"' % etag.replace('\\', '\\\\').replace('"', '\\"')
 
+if sys.version_info >= (2, 6):
+    def same_origin(url1, url2):
+        """
+        Checks if two URLs are 'same-origin'
+        """
+        p1, p2 = urlparse.urlparse(url1), urlparse.urlparse(url2)
+        return (p1.scheme, p1.hostname, p1.port) == (p2.scheme, p2.hostname, p2.port)
+else:
+    # Python 2.4, 2.5 compatibility. This actually works for Python 2.6 and
+    # above, but the above definition is much more obviously correct and so is
+    # preferred going forward.
+    def same_origin(url1, url2):
+        """
+        Checks if two URLs are 'same-origin'
+        """
+        p1, p2 = urlparse.urlparse(url1), urlparse.urlparse(url2)
+        return p1[0:2] == p2[0:2]