Adding back the #wait_until to Capybara v2+ gem. Can probably be adapted for cucumber tests.

capybara_wait_until.rb

# add this file capybara_wait_until.rb to your /test directory

module Capybara
  class Session
    ##
    #
    # Retry executing the block until a truthy result is returned or the timeout time is exceeded
    #
    # @param [Integer] timeout   The amount of seconds to retry executing the given block
    #
    # this method was removed in Capybara v2 so adding it back if not already defined
    #
    unless defined?(wait_until)
      def wait_until(timeout = Capybara.default_wait_time)
        Capybara.send(:timeout, timeout, driver) { yield }
      end
    end
  end
end

# Use it this way in /test/integration_test_helper.rb:

require 'capybara_wait_until'
module ActionController
  class IntegrationTest
    def wait_for_ajax_to_complete
      # pass timeout in seconds if you need to override default_wait_time
      page.wait_until { page.evaluate_script('jQuery.active === 0') }
    end
  end
end

I really, really love Capybara and I also love Ruby for the ability to veto gem authors' choices. ;)

I really, really love Capybara and I also love Ruby for the ability to veto gem authors' choices. ;)

I agree, I need something like this as well for a very special case of having to wait on Resque workers to finish processing before it updates the page via a push notification. Capybara just doesn't handle this 'special' case at all.

I agree, I need something like this as well for a very special case of having to wait on Resque workers to finish processing before it updates the page via a push notification. Capybara just doesn't handle this 'special' case at all.

Why can't you use #synchronize? http://rubydoc.info/github/jnicklas/capybara/master/Capybara/Node/Base#synchronize-instance_method

Why can't you use #synchronize? http://rubydoc.info/github/jnicklas/capybara/master/Capybara/Node/Base#synchronize-instance_method

When you use find, capybara will be actually waiting for the element to appear (by using synchronize behind the scenes), so there's little reason to use wait_for here.

Moreover, this couples your acceptance tests to your underlying implementation and will be a pain to deal with when you decide to use websockets there, or just remove ajax calls for some reason.

When you use find, capybara will be actually waiting for the element to appear (by using synchronize behind the scenes), so there's little reason to use wait_for here.

Moreover, this couples your acceptance tests to your underlying implementation and will be a pain to deal with when you decide to use websockets there, or just remove ajax calls for some reason.

@josepjaume - can you give a code example on using synchronize? I get an error that the method doesn't exist.

@josepjaume - can you give a code example on using synchronize? I get an error that the method doesn't exist.

When I try to call wait_until, I get:
class or module required for rescue clause

When I try to call wait_until, I get:
class or module required for rescue clause

to be more specific:

Failure/Error: page.wait_until {
     TypeError:
       class or module required for rescue clause
     # ./spec/support/capybara/wait_until.rb:13:in `wait_until'

to be more specific:

Failure/Error: page.wait_until {
     TypeError:
       class or module required for rescue clause
     # ./spec/support/capybara/wait_until.rb:13:in `wait_until'