Search code examples
selenium-webdrivertestingrspeccapybararspec-rails

Strange problem when running system tests as a whole with RSpec and Capybara/Selenium

I've regular RSpec Rails system tests using Capybara/Selenium, like this:

RSpec.describe "Model", type: :system do
  context "update model info1" do
    # ...
  end

  context "update model info2" do
    # ...
  end

  context "update model info3" do
    # ...
  end

  # ...
end

If I run these tests individually (e.g. context "update model info1" or context "update model info2") then each of them passes with success.

However, if I run them as a whole from the "root" level (i.e. from RSpec.describe "Model") then some of them fails.

Tests always run sequentially.

What is the problem? How can I make tests to pass with success even when they run as a whole?

If it can be useful, I'm running tests on MacOs Ventura 13.0 with rails 7.0.4.3, rspec-rails 6.0.1, capybara 3.38.0, and selenium-webdriver 4.8.1.

Solution

  • Problems occur because the tested CSS-JS widgets have animations and transitions, which require time... and because my tests do not wait for that time.

    The fix is to set the Capybara.disable_animation option to true:

    Capybara.disable_animation = true

    See also this.