Domain whitelist

gistfile1.py

import re

class DomainWhitelist(object):
    url_re = re.compile(
        r'([a-z0-9]+([\-\.]{1}[a-z0-9]+)*\.[a-z]{2,6})',
        re.IGNORECASE)

    kogama_re = re.compile(
        r'^(http[s]?://)?([a-z0-9-.]{2,8})?kogama.com(.br)?',
        re.IGNORECASE)

    def check(self, value):
        for url in self.url_re.findall(value):
            if not self.kogama_re.match(''.join(url)):
                return False
        return True

@simonz05 take a look: http://rubular.com/r/LFWxz7kpgd

I just changed the "dot" part:

r'^(http[s]?://)?([a-z0-9]{2,8}.)?kogama.com(.br)?'

The modified reg-ex is for testing if a link is in the white-list or not. It's the first reg-ex that determines if it's a link or not, and thus it's the first reg-ex that need to handle all URL-cases. This is the one: r'http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*(),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+'

Got it.

This one that you send is not matching www.domain.com without http|s: http://rubular.com/r/FqYasTLz5l

This one is: http://rubular.com/r/FAsg1mTTnB

r'(http[s]?://|www)(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*\(\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+'

I think I'll simplify it to just look for domain names. It should solve the issue.

[a-z0-9]+([\-\.]{1}[a-z0-9]+)*\.[a-z]{2,6}

http://rubular.com/r/GOuWVw98Qi

We may have problems with usernames don't we? http://rubular.com/r/nFfQpESoPP

Yes and no. You wont be able to write a comment with the username "simon.kogama" in it. However, that is not something we can solve, since domain and user names follow some of the same rules.

We do however not run these validations on usernames directly. Usernames are upon creation checked versus a much stricter regex which limits the possibilities. You can create a username like this xxx.com, but that is about it.