Capybara-webkit EOF error and server crashing when mocking with puffing-billy

[lukegrecki]

I am currently using capybara-webkit (1.11.1) with puffing_billy (0.8.0) to make mock requests from javascript, and am consistently running into the following error:

1) Analytics Dashboard when the user visits analytics dashboard today's summary details are visible
     Got 0 failures and 2 other errors:

     1.1) Failure/Error: visit analytics_dashboard_path

          EOFError:
            end of file reached
          # <internal:prelude>:76:in `__read_nonblock'
          # <internal:prelude>:76:in `read_nonblock'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/connection.rb:41:in `read'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/connection.rb:32:in `gets'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/browser.rb:299:in `check'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/browser.rb:211:in `command'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/browser.rb:19:in `visit'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/driver.rb:50:in `visit'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-2.7.1/lib/capybara/session.rb:233:in `visit'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-2.7.1/lib/capybara/dsl.rb:52:in `block (2 levels) in <module:DSL>'
          # ./spec/features/analytics/dashboard_integration_spec.rb:139:in `load_analytics_dashboard'
          # ./spec/features/analytics/dashboard_integration_spec.rb:276:in `block (3 levels) in <top (required)>'
          # ./spec/ui_spec_helper.rb:53:in `block (2 levels) in <top (required)>'

     1.2) Failure/Error: example.run

          Capybara::Webkit::CrashError:
            The webkit_server process crashed!

              Connection reset by peer

            This is a bug in capybara-webkit. For help with this crash, please visit:

            https://github.com/thoughtbot/capybara-webkit/wiki/Reporting-Crashes
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/browser.rb:215:in `rescue in command'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/browser.rb:205:in `command'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/browser.rb:79:in `current_url'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-webkit-1.11.1/lib/capybara/webkit/driver.rb:46:in `current_url'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-2.7.1/lib/capybara/session.rb:181:in `current_url'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-screenshot-1.0.13/lib/capybara-screenshot/rspec.rb:55:in `block in after_failed_example'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-2.7.1/lib/capybara.rb:352:in `using_session'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-screenshot-1.0.13/lib/capybara-screenshot/rspec.rb:54:in `after_failed_example'
          # /Users/lukegrecki/.rvm/gems/ruby-2.3.1@backoffice/gems/capybara-screenshot-1.0.13/lib/capybara-screenshot/rspec.rb:84:in `block (2 levels) in <top (required)>'
          # ./spec/ui_spec_helper.rb:53:in `block (2 levels) in <top (required)>'
          # ------------------
          # --- Caused by: ---
          # Errno::ECONNRESET:
          #   Connection reset by peer
          #   <internal:prelude>:76:in `__read_nonblock'

To give a little more context I'm stubbing about 8 requests total and these may be coming in rather fast. The first set of 4 requests are options requests, and the second set of 4 requests have the same urls as the first but are get requests asking for data.

I wasn't sure if this was a capybara-webkit issue or a puffing-billy one so I also opened an issue in the former repo.

I know puffing-billy doesn't parallelize the processing of mock requests. Could too many ajax requests coming into capybara-webkit + puffing-billy cause a crash? Any ideas how to work around this?

I appreciate any help or direction.

[ronwsmith]

I haven't heard of something like this before.

Does everything work as expected using webkit without puffing-billy?

[lukegrecki]

This turned out to be an issue locally for me (still unsolved) but it passes locally consistently for someone else on the team.