Now, more more ..
create issue later.
It seems like turnip is completely different.
I wrote about RSpec support policy and turnip release plan. Please check it! https://github.com/jnicklas/turnip/issues/158#issuecomment-119049054
Please check the changelog https://github.com/jnicklas/turnip/releases/tag/v2.0.0
Thanks :pizza:
I plan to release 3.0.0 and I've released 3.0.0.beta.2 https://github.com/jnicklas/turnip/releases/tag/v3.0.0.pre.beta.2 :muscle:
Please try 3.0.0.beta2 in your code if you are free :bow:
hmm ok it seems that when capybara is clicking the link, it's actually going to the destination of the link, and not doing the ajax thing
the link being clicked looks like:
<a class="pb-1 border-bottom border-light relevance-label-control" id="55-relevant-label-control" data-remote="true" rel="nofollow" data-method="patch" href="/relevance_labels/55?filter=1&project=1&relevance=true&user=1"><i aria-hidden="true" class="fa-lg fa-thumbs-up grey-text far" data-html="true" data-toggle="tooltip" id="55-relevant-label-icon" title="" data-original-title="Label this article as <strong>relevant</strong> to this stream"></i></a>
so, it's
data-remote and patch and definitely uses ajax in the regular browser. it just doesn't seem like capybara is using the selenium driver
Here's the feature:
@elasticsearch @javascript
Scenario Outline: Labeling articles
When a specific stream is viewed
And the label <new_label> is applied to an article whose label is <current_label>
Then only the label <new_label> is indicated
And a message about <new_label> is displayed
Examples:
| current_label | new_label |
| unlabeled | relevant |
the
@elasticsearch part definitely works
that should probably be selenium
and not javascript
hmm no change
so, it appears that the
@javacript / @selenium are being ignored
ok, getting further. here's my turnip_helper.rb now:
require 'rails_helper'
require 'turnip/capybara' # to use Capybara DSL methods in steps
require 'capybara'
require 'webdrivers'
Dir.glob("spec/steps/**/*steps.rb") { |f| load f, true }
Capybara.current_driver = :selenium
Capybara.javascript_driver = :seleniumBut now when I run the feature, I get the following error:
Failure/Error: visit(project_path(@labeling_project.slug, f: @labeling_stream.slug, q: @labeling_stream.query))
Selenium::WebDriver::Error::UnknownError:
invalid argument: can't kill an exited process
ok, i'm running this in a container, and found SeleniumHQ/selenium#6622, which implies running headless firefox, and now I've made progress. Finally at the point of getting a failed test, but for a weird reason
tl;dr: now this is pure user error and nothing to do with turnip
ok, now I am having an issue with the "Wrong" javascript driver being run...
require 'rails_helper'
require 'turnip/capybara' # to use Capybara DSL methods in steps
require 'capybara'
require 'webdrivers'
Dir.glob("spec/steps/**/*steps.rb") { |f| load f, true }
Capybara.register_driver :firefox_headless do |app|
options = ::Selenium::WebDriver::Firefox::Options.new
options.args << '--headless'
Capybara::Selenium::Driver.new(app, browser: :firefox, options: options)
end
Capybara.javascript_driver = :firefox_headless
Capybara.current_driver = :firefox_headlessWhen I get down into a test and use a binding.pry, I get the following:
[1] pry(#<RSpec::ExampleGroups::UsersCanManageRelevanceLabelsOnArticles::LabelingArticles>)> Capybara.current_driver
=> :rack_test
[2] pry(#<RSpec::ExampleGroups::UsersCanManageRelevanceLabelsOnArticles::LabelingArticles>)> Capybara.javascript_driver
=> :firefox_headless
This message was deleted
something seems like it's resetting the javascript driver to
:rack_test and I don't know why
on to the next problem