We accidentally removed the `count` option in commit 55fb14ac, which
made the translation return a hash.
The test is a bit hacky, which makes me think changing the user
interface would probably be a better solution.
After adding investment translatable fields to forms, they will be
generated with nested translations names and ids so we can no longer
use standard id locator.
Using input label text to fill in fields works with both types of forms.
The `type: :feature` is automatically detected by RSpec because these
tests are inside the `spec/features` folder. Using `feature` re-adds a
`type: :feature` to these files, which will result in a conflict when we
upgrade to Rails 5.1's system tests.
Because of this change, we also need to change `background` to `before`
or else these tests will fail.
We were expecting the page not to have content which is actually there.
The test passed (most of the time) because before clicking the
"Milestones" link the content was not present, and we checked the page
content before the AJAX request generated by clicking the link had
finished.
We were sending a new request without checking the previous one had
finished, and if it hadn't finished, the test failed.
This test started to fail after upgrading to Rails 5, since we removed
the change done in commit eda47eff which set `config.allow_concurrency`
to `false` in the test environment.
While we could re-introduce that configuration option, which might have
side effects, an easier solution is to check an AJAX request has been
completed before sending a new request depending on the previous one
seems to be a more solid option.
Note this test failed with two possible errors: "undefined method
`heading' for nil:NilClass" and "stale element reference: element is not
attached to the page document". This change fixes the second error; it
might fix the first error as well, but since I couldn't reproduce it
locally, we'll only be sure when this test stops failing in travis
builds.
There was a typo: `new_order = eq(all(` instead of `new_order = all(`,
which was causing the tests to pass.
However, the final expectation should test that we keep the same order
in the same session, and we were accidentally testing the opposite.
We're also adding an extra check to verify there are investments on the
page, since in some cases we were accessing pages with no investments,
and so these tests were always passing.
This is the actual number of investments per page in the index action.
Also note one test was generating 100 extra records, which made the test
take more than 40 seconds (on my machine).
We were showing only the ones being shown in the current page because
we were modifying `@investments` using a method which used
`@investments`, and we were calling that method twice.
There are many possible solutions: using a local variable to store the
result of the `investments` method, modifying `@investments` after
modifying `@investments_map_coordinates`, ... I've used the one which in
my humble opinion is a bit less fragile: not using `@investments` inside
the `investments` method. That way, the `investments` method will always
return the same result.
Note `stub_const("Budgets::InvestmentsController::PER_PAGE", 2)`
wouldn't work because `Budgets::InvestmentsController` isn't loaded when
that line is executed. So we need to load it. Instead of requiring the
file, using `#{Budgets::InvestmentsController}` seems to be an easier
solution.
We were filtering by winners investments for finished budget without
having in consideration other filters.
We want the default filter to be `winners` for finished budgets.
