Ticket #9764: idn_support.diff

django/forms/fields.py

@@ -432 +432 @@
         RegexField.__init__(self, email_re, max_length, min_length, *args,
                             **kwargs)
 
+    def clean(self, value):
+        try:
+            value = super(EmailField, self).clean(value)
+        except ValidationError:
+            # Trivial case failed. Try for possible IDN domain-part
+            if value and u'@' in value:
+                parts = value.split(u'@')
+                domain_part = parts[-1]
+                try:
+                    parts[-1] = parts[-1].encode('idna')
+                except UnicodeError:
+                    pass # invalid domain part
+                # After validation revert ACE encoded domain-part to 
+                # original (IDN) value as suggested by RFC 3490
+                value = super(EmailField, self).clean(u'@'.join(parts)
+                    ).replace(parts[-1], domain_part)
+            else:
+                raise
+        return value
+
 try:
     from django.conf import settings
     URL_VALIDATOR_USER_AGENT = settings.URL_VALIDATOR_USER_AGENT
@@ -547 +567 @@
         self.user_agent = validator_user_agent
 
     def clean(self, value):
+        if value:
+            value = smart_unicode(value)
         # If no URL scheme given, assume http://
         if value and '://' not in value:
             value = u'http://%s' % value
-        # If no URL path given, assume /
-        if value and not urlparse.urlsplit(value)[2]:
-            value += '/'
+        if value:
+            splitted = urlparse.urlsplit(value)
+            # IDN -> ACE
+            try:
+                netloc_ace = splitted[1].encode('idna')
+            except UnicodeError: # invalid domain part
+                netloc_ace = splitted[1]
+            value = value.replace(splitted[1], netloc_ace)
+            # If no URL path given, assume /
+            if not splitted[2]:
+                value += u'/'
         value = super(URLField, self).clean(value)
         if value == u'':
             return value
+        # ACE -> IDN
+        value = value.replace(netloc_ace, splitted[1])
         if self.verify_exists:
             import urllib2
             headers = {

tests/regressiontests/forms/fields.py

@@ -733 +733 @@
 ValidationError: [u'This field is required.']
 >>> f.clean('person@example.com')
 u'person@example.com'
+
+Address with IDN domain-part:
+>>> f.clean('local@domain.with.idn.xyzäöüßabc.part.com')
+u'local@domain.with.idn.xyz\xe4\xf6\xfc\xdfabc.part.com'
+
 >>> f.clean('foo')
 Traceback (most recent call last):
 ...
@@ -765 +770 @@
 Traceback (most recent call last):
 ...
 ValidationError: [u'Enter a valid e-mail address.']
+>>> f.clean('foo@.com')
+Traceback (most recent call last):
+...
+ValidationError: [u'Enter a valid e-mail address.']
 
 EmailField also access min_length and max_length parameters, for convenience.
 >>> f = EmailField(min_length=10, max_length=15)
@@ -868 +877 @@
 u'http://200.8.9.10/'
 >>> f.clean('http://200.8.9.10:8000/test')
 u'http://200.8.9.10:8000/test'
+
+IDN domain
+>>> f.clean('http://some.idn.xyzäöüßabc.domain.com:123/blah')
+u'http://some.idn.xyz\xe4\xf6\xfc\xdfabc.domain.com:123/blah'
+
 >>> f.clean('foo')
 Traceback (most recent call last):
 ...