Question

I’ve written a regex to help validate a String for game character names. It’s somehow passing seemingly invalid strings and not passing seemingly valid strings.

Requirements:

• Starts with a capital letter
• Has any number of alphanumeric characters after that (this includes spaces)

This is the rails code that does the validation in the Character Model:

validates :name, format: { with: %r{[A-Z][a-zA-Z0-9\s]*} }

Here’s the unit test I’m using

test "character name should be properly formatted and does not contain any special characters" do
    character = get_valid_character
    assert character.valid?

    character.name = "aBcd"
    assert character.invalid?, "#{character.name} should be invalid"

    character.name = "Number 1"
    assert character.valid?, "#{character.name} should be valid"

    character.name = "McDonalds"
    assert character.valid?, "#{character.name} should be valid"

    character.name = "Abcd."
    assert character.invalid?, "#{character.name} should be invalid"

    character.name = "Abcd%"
    assert character.invalid?, "#{character.name} should be invalid"
end

The problems:
The regex passes “aBcd”, “Abcd.”, and “Abcd%” when it shouldn’t. Now, I know this works because I tested this out in Python and it works just as you would expect.

What gives?

Thank you for your help!

Answer 1

Regular expressions look for matches anywhere in the given string unless told otherwise.

So the test string 'aBcd' is invalid, but it contains a valid substring: 'Bcd'. Same with 'Abcd%', where the valid substring is 'Abcd'.

If you want to match the entire string, use this as your regex:

# \A matches string beginning, \z matches string end
%r{\A[A-Z][a-zA-Z0-9\s]*\z}

PS: Some people will say to match the beginning of a string with ^ and the end with $. In Ruby, those symbols match the beginning and end of a line, not a string. So "ABCD\n%" would still match if you used ^ and $, but won’t match if you use \A and \z. See the Rails security guide for more on this.