Capybara: a better way to check if text is visible

Better version

### IMPORTANT NOTE:  See the comment thread below for a more concise way to determine this now, using has_css...


#A better 'I should not see' for Capybara that lets jQuery determine visibility of the text you're looking for.

Then /^"([^\"]*)" should not be visible$/ do |text|
  finder_script = %{
    function is_text_visible_on_page(text) {
      var match = false;
      $('*:visible')
        .contents()
        .filter(function() {
          //collect text nodes
          return this.nodeType === 3;
        })
        .each(function() {
          if (this.textContent.indexOf(text) != -1) {
            match = true;
            return false;
          }
        });
      return match;
    }

    is_text_visible_on_page('#{text}');
  }

  assert ! page.driver.evaluate_script(finder_script)
end

The logic looks reversed in this example.

@gerryster Possibly so. This is an old script and I'm not using it anymore. These days I'm using:

Then /^"([^"]*)" should not be visible$/ do |text| begin assert page.find(text).visible? != true rescue Capybara::ElementNotFound end end

I'm guessing you're using something even newer now since that version doesn't work? :)

@ctide: yep. These days I'm not using cucumber at all, and just using rspec with capybara directly, like:

page.should have_no_css('a', :text => 'foo', :visible => true)

+1, this works very well. Confused me at first, but I scoped this within the location, then if the thing exists use the above line to make sure the "page has no css with that text that's visible."

Wound up making a helper like this:

  def restaurant_should_be_hidden(location, text)
    within(location) do
      if page.has_content?(text)
        page.should have_no_css('a', :text => text, :visible => true)
      end    
    end
  end