Regex retrieved from a MYSQL database to validate email addresses using Ruby on Rails

Question

I am using Ruby on Rails 3 and a MYSQL database. I would like to retrieve a regex from the database and then use that value to validate email addresses.

I aim to not put the regex value in line in my RoR application code, but outside so that the value can be recalled for other usages and from other places.

In order to populate the database, I put in my 'RAILS_ROOT/db/seed.rb' the following:

Parameter.find_or_create(
  :param_name        => 'email_regex',
  :param_value       => "[a-z0-9!#\$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#\$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?",
)

Notice : in the 'seed.rb' file I edited a little bit the original regex from www.regular-expressions.info adding two \\ just before $ . Here it is the difference:

#original from www.regular-expressions.info
[a-z0-9!#$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?
#edited by me
[a-z0-9!#\$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#\$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?

After run rake db:seed in the Terminal, in MYSQL database I have this value ( without \\ near $ ):

[a-z0-9!#$%&'*+/=?^_`{|}~-]+(?:.[a-z0-9!#$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?

Then in my RoR application I use the regex this way:

def validate(string)
  email_regex = Regexp.new(Parameter.find_by_param_name('email_regex').param_value)
  if email_regex.match(string)
    return true
  else
    return false
  end
end

The problem using the above regex is that I can successfully validate also email addresses with double '@' or without the final part like these:

name@surname@gmail.com # Note the double '@'
test@gmail

Of course I would like to refuse those email addresses . So, how can I adjust that? Or, how can I get what I want?

---

I tried also to seed these regex:

#case 1
\A[a-z0-9!#\$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#\$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\Z

#case 2
\\A[a-z0-9!#\$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#\$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\\Z

#case 3
^[a-z0-9!#\$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#\$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?$

that in the MYSQL database become respectively:

#case 1
A[a-z0-9!#$%&'*+/=?^_`{|}~-]+(?:.[a-z0-9!#$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?Z

#case 2
\A[a-z0-9!#\$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#\$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\Z

#case 3
^[a-z0-9!#$%&'*+/=?^_`{|}~-]+(?:.[a-z0-9!#$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?$

but also them don't work as expected.

---

UPDATE

Debugging I have

--- !ruby/regexp /[a-z0-9!#$%&'*+\/=?^_`{|}~-]+(?:.[a-z0-9!#$%&'*+\/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?/

that means that just before of all / characters Ruby added a \\ character. Can be that my problem? In the 'seed.rb' file I tryed to escape all / adding \\ statements but the debug output is always the same.

Answer 1

There are so many things wrong on so many levels here…

1. Storing application configuration in your database isn't recommended; slower performance, potential catch 22s (like how do you configure your database, from your database), etc. Try something like SettingsLogic if you don't want to have to build your own singleton configuration or use an initializer.
2. Rails has built in validation functionality as a mixin that's automatically part of any models inheriting from ActiveRecord::Base. You should use it, rather than define your own validation routines, especially for basic cases like this.
3. You can actually have an email address with multiple @ signs, provided the first is escaped with a backslash or the local portion of the address is quoted.
4. Why are you escaping $ characters in a character class where they have no special meaning?
5. Regular expressions are okay for a very basic validation of an email address to make sure you didn't get complete garbage data to pass off to your mail server, but they aren't the best way to verify an email address.

I suggest you have a good look at the validations guide at RubyOnRails.org.

Answer 2

You shouldn't reinvent this wheel. See http://lindsaar.net/2010/1/31/validates_rails_3_awesome_is_true for a standard way to validate email addresses in Rails 3.

If you do choose to reinvent the wheel, don't use a regular expression. The gory details of why this is a bad idea are explained in http://oreilly.com/catalog/9780596528126 , along with a very, very complicated regular expression that almost does it.