Question

I’m looking at some RSpec code in Chapter 6 of Michael Hartl’s [Rail Tutorial]: http://ruby.railstutorial.org/chapters/modeling-users#code:validates_uniqueness_of_email_test. It was written to test the uniqueness validation method of the email attribute of the user model. It looks like this:

Listing 6.18. A test for the rejection of duplicate email addresses. (*spec/models/user_spec.rb*)

require 'spec_helper'

describe User do

  before do
    @user = User.new(name: "Example User", email: "[e-mail]")
  end
  subject { @user }
  .
  .
  .
  describe "when email address is already taken" do
    before do
      user_with_same_email = @user.dup
      user_with_same_email.save
    end

    it { should_not be_valid }
  end
end

And here is the user model validation code:

validates :email, presence: true, format: { with: VALID_EMAIL_REGEX },
                    uniqueness: true

My question is: How does the validating actually work? There is a user with an identical email address already saved into the database and the Rails validation on the @user instance comes back as invalid before I’ve even tried to save it? Does the Rails validator use current data to validate instances of the user even if they are just being stored in memory and not currently attempting to be added to the database?

Answer 1

be_valid calls valid? on @user. It’s how rspec works. And valid? calls run_validations! which in turn calls run_callbacks :validate. You can follow the methods chain through source code: valid? is defined in activerecord / lib / active_record / validations.rb and it calls super valid? in activemodel / lib / active_model / validations.rb

That’s about how it works. And the model can be valid or not even if it is unsaved. If it was never saved (new_record? returns true) as in your case validations on :create are called. Basically if model is valid it can be saved. If it was saved (new_record? returns false but it can contain unsaved changes) validations on :update are called.