In a recent post, Steve Klabnik argued that when writing Cucumber steps, we should base our selectors on the “rel” attribute of a link rather than using the traditional selectors (:id, :class, text, etc.). He argues that IDs shouldn’t be used because they are unique and don’t provide a general enough solution, and class attributes have everything to do with styles and nothing to do with semantics and data. His solution, as I’ve already mentioned, is to use an anchor tag’s “rel” attribute.
At the time of his writing, I also happened across Capybara’s “add_selector” method and was looking for an opportunity to use it. Steve’s article made for a perfect opportunity.
If you don’t already know, “::add_selector” “[Add’s] a new selector to Capybara” (shocking, I know). By using this method, you can simplify the way you find elements. So rather than using this:
num = 3
page.find(:xpath, ".//table/tr[#{num}]")
You might use something like this:
page.find(:table_row, num)
Back to Klabnik’s post. Like I said, Steve thinks we should be looking for links based on their “rel” attribute. To do this in Capybara we normally have to do something like this:
find("//a[@rel='edit-article']").click
But rather than using this bit of xpath nastiness, let’s use Capybara’s “::add_selector” method.
I’ve created a “selectors.rb” file in my “Rails.root/features/support” directory and added the following finder:
Capybara.add_selector(:link) do
xpath {|rel| ".//a[@rel='#{rel}']"}
end
Now, I can use this in my steps:
find(:link, "edit-article")
Not only is this easier to remember, but it reads better too.