Obsolete nodes are not refreshed when using Capybara's #within and/or #synchronize #242

shepmaster · 2023-09-19T14:28:40Z

Reproduction

This code uses Capybara's synchronize inside of within. The code inside synchronize always fails, but that's a simplification from the real application. The page has JavaScript that fully replaces the #parent element that contains the#child element, which is what Capybara is looking for.

Gemfile

# frozen_string_literal: true

source "https://rubygems.org"

gem "cuprite", "~> 0.14.3"
gem "capybara", "~> 3.39"

repro.rb

require 'capybara'
require 'capybara/dsl'
require 'capybara/cuprite'

include Capybara::DSL

Capybara.javascript_driver = :cuprite
Capybara.register_driver(:cuprite) do |app|
  Capybara::Cuprite::Driver.new(app)
end

Capybara.current_driver = :cuprite
Capybara.app_host = 'http://[::1]:8888'
Capybara.run_server = false

class DummyError < StandardError; end

visit('/')

within('#parent') do
  ok_errors = page.driver.invalid_element_errors + [DummyError]
  page.document.synchronize(Capybara.default_max_wait_time, errors: ok_errors) do
    page.all('#child', count: 1)

    # We always fail this test. If everything is working correctly,
    # this program should report this exception.
    raise DummyError
  end
end

index.html

<!doctype html>
<html>
  <head>
    <title>Reproduction</title>
  </head>

  <body>
    <div id="parent">
      <div id="child">Child 1</div>
    </div>
  </body>

  <script type="text/javascript">
    function sleep(ms) {
      return new Promise(resolve => setTimeout(resolve, ms));
    }

    async function main() {
      await sleep(1000);

      const parent = document.getElementById('parent');
      const body = parent.parentElement;
      parent.remove();

      const newParent = document.createElement('div');
      newParent.id = 'parent';

      const newChild =  document.createElement('div');
      newChild.id = 'child';
      newChild.innerText = 'Child 2';

      newParent.appendChild(newChild);
      body.appendChild(newParent);
    }

    main()
  </script>
</html>

Steps

Serve the page via python3 -m http.server --bind '::1' 8888
Run the test via bundle exec ruby repro.rb

Output

gems/cuprite-0.14.3/lib/capybara/cuprite/node.rb:23:in `rescue in command': The element you are trying to interact with is either not part of the DOM, or is not currently visible on the page (perhaps display: none is set). It is possible the element has been replaced by another element and you meant to interact with the new element. If so you need to do a new find in order to get a reference to the new element. (Capybara::Cuprite::ObsoleteNode)
	from gems/cuprite-0.14.3/lib/capybara/cuprite/node.rb:20:in `command'
	from gems/cuprite-0.14.3/lib/capybara/cuprite/node.rb:48:in `find'
	from gems/cuprite-0.14.3/lib/capybara/cuprite/node.rb:44:in `find_css'
	from gems/capybara-3.39.2/lib/capybara/node/base.rb:108:in `find_css'
	from gems/capybara-3.39.2/lib/capybara/queries/selector_query.rb:253:in `find_nodes_by_selector_format'
	from gems/capybara-3.39.2/lib/capybara/queries/selector_query.rb:166:in `block in resolve_for'
	from gems/capybara-3.39.2/lib/capybara/node/base.rb:77:in `synchronize'
	from gems/capybara-3.39.2/lib/capybara/queries/selector_query.rb:165:in `resolve_for'
	from gems/capybara-3.39.2/lib/capybara/node/finders.rb:265:in `block in all'
	from gems/capybara-3.39.2/lib/capybara/node/base.rb:77:in `synchronize'
	from gems/capybara-3.39.2/lib/capybara/node/finders.rb:264:in `all'
	from gems/capybara-3.39.2/lib/capybara/session.rb:773:in `all'
	from repro.rb:25:in `block (2 levels) in <main>'
	from gems/capybara-3.39.2/lib/capybara/node/base.rb:84:in `synchronize'
	from repro.rb:24:in `block in <main>'
	from gems/capybara-3.39.2/lib/capybara/session.rb:365:in `within'
	from gems/capybara-3.39.2/lib/capybara/dsl.rb:52:in `call'
	from gems/capybara-3.39.2/lib/capybara/dsl.rb:52:in `within'
	from repro.rb:20:in `<main>'
gems/ferrum-0.13/lib/ferrum/browser/client.rb:88:in `raise_browser_error': No node with given id found (Ferrum::NodeNotFoundError)
	from gems/ferrum-0.13/lib/ferrum/browser/client.rb:49:in `command'
	from gems/ferrum-0.13/lib/ferrum/page.rb:276:in `command'
	from gems/cuprite-0.14.3/lib/capybara/cuprite/browser.rb:74:in `find_within'
	from gems/cuprite-0.14.3/lib/capybara/cuprite/node.rb:21:in `command'
	from gems/cuprite-0.14.3/lib/capybara/cuprite/node.rb:48:in `find'
	from gems/cuprite-0.14.3/lib/capybara/cuprite/node.rb:44:in `find_css'
	from gems/capybara-3.39.2/lib/capybara/node/base.rb:108:in `find_css'
	from gems/capybara-3.39.2/lib/capybara/queries/selector_query.rb:253:in `find_nodes_by_selector_format'
	from gems/capybara-3.39.2/lib/capybara/queries/selector_query.rb:166:in `block in resolve_for'
	from gems/capybara-3.39.2/lib/capybara/node/base.rb:77:in `synchronize'
	from gems/capybara-3.39.2/lib/capybara/queries/selector_query.rb:165:in `resolve_for'
	from gems/capybara-3.39.2/lib/capybara/node/finders.rb:265:in `block in all'
	from gems/capybara-3.39.2/lib/capybara/node/base.rb:77:in `synchronize'
	from gems/capybara-3.39.2/lib/capybara/node/finders.rb:264:in `all'
	from gems/capybara-3.39.2/lib/capybara/session.rb:773:in `all'
	from repro.rb:25:in `block (2 levels) in <main>'
	from gems/capybara-3.39.2/lib/capybara/node/base.rb:84:in `synchronize'
	from repro.rb:24:in `block in <main>'
	from gems/capybara-3.39.2/lib/capybara/session.rb:365:in `within'
	from gems/capybara-3.39.2/lib/capybara/dsl.rb:52:in `call'
	from gems/capybara-3.39.2/lib/capybara/dsl.rb:52:in `within'
	from repro.rb:20:in `<main>'

Observations

If I add puts page.has_selector?('#child', text: 'Child 2') before the synchronize block, it will print out true and the program will (correctly) exit with DummyError.
If I replace cuprite with apparition, the program will (correctly) exit with DummyError.

The text was updated successfully, but these errors were encountered:

infernalmaster · 2023-11-03T13:50:59Z

I did some monkey-patching to fix similar issue

gem 'capybara', '3.37.1'
gem 'cuprite', '0.14.3'

module NodePatch
  def command(name, *args)
    browser.send(name, node, *args)
  rescue Ferrum::NodeNotFoundError => e
    # Throw ElementNotFound because it's not inside #invalid_element_errors
    # so it would not be suppresed inside Capybara::Queries::SelectorQuery#matches_filters?
    # and then error would force element to reload Capybara::Node::Element#reload
    # instead of just ignore it
    #
    # Original code:
    # raise ObsoleteNode.new(self, e.response)
    raise Capybara::ElementNotFound, e.response
  rescue Ferrum::BrowserError => e
    case e.message
    when 'Cuprite.MouseEventFailed'
      raise MouseEventFailed.new(self, e.response)
    else
      raise
    end
  end
end

Capybara::Cuprite::Node.prepend NodePatch

I still think that there should be better way to do that

baygeldin · 2024-02-02T14:32:59Z

I spent a lot of time debugging Capybara/Cuprite/Ferrum source code trying to get to the root of this issue just to realize it's not actually a bug. There is a good reason outlined by @jnicklas in this comment. Basically, when you use #all and the node on a page becomes obsolete, you can't be sure that if you reload results of #all they would be the same. However, it is possible to make #all results reload, just pass allow_reload option (e.g. page.all('#child', count: 1, allow_reload: true)).

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Obsolete nodes are not refreshed when using Capybara's #within and/or #synchronize #242

Obsolete nodes are not refreshed when using Capybara's #within and/or #synchronize #242

shepmaster commented Sep 19, 2023

infernalmaster commented Nov 3, 2023

baygeldin commented Feb 2, 2024

Obsolete nodes are not refreshed when using Capybara's #within and/or #synchronize #242

Obsolete nodes are not refreshed when using Capybara's #within and/or #synchronize #242

Comments

shepmaster commented Sep 19, 2023

Reproduction

Gemfile

repro.rb

index.html

Steps

Output

Observations

infernalmaster commented Nov 3, 2023

baygeldin commented Feb 2, 2024