Issue 2234 - Add a WSGI controller to collect email addresses for the Adblock Browser iOS launch

sitescripts/utils.py

Index: sitescripts/utils.py
===================================================================
--- a/sitescripts/utils.py
+++ b/sitescripts/utils.py
@@ -113,6 +113,25 @@
   else:
     subprocess.Popen([config.get('DEFAULT', 'mailer'), '-t'], stdin=subprocess.PIPE).communicate(mail.encode('utf-8'))
 
+def encode_email_address(email):
+  '''
+  Validates and encodes an email address.
+
+  The validation implemented here is very rudamentery and not meant
+  to be complete, as full email validation can get extremly complicated
+  and is rarely needed. This function is primarily making sure that the
+  email address contains no whitespaces and only valid ASCII characters.
+  '''
+  match = re.search(r'^([^@\s]+)@([^@\s]+)$', email)
+  if not match:
+    raise ValueError
+
+  try:
+    return email.encode('ascii')
+  except UnicodeEncodeError:
+    return '%s@%s' % (match.group(1).encode('ascii'),
+                      match.group(2).encode('idna'))

[Wladimir Palant, 2015/04/23 16:04:40]

This seems to have two issues:

1. match.group(1).encode('ascii') can fail and will produce another
UnicodeEncodeError - yet the caller is only expecting ValueError.

2. I'm unsure whether match.group(2).encode('idna') will work - it requires
unicode, not str. What do we have here?

[Sebastian Noack, 2015/04/23 16:29:41]

issubclass(UnicodeEncodeError, ValueError) => True
Sure, a function supposed to encode text, is supposed to be called with a
unicode object, as we do here.

+
 _template_cache = {}
 
 def get_template(template, autoescape=True):