RSpec - A JSON text must at least contain two octets - ruby-on-rails

Question

Two failing specs after upgrading rspec-rails (2.5.2 -> 3.8.1) and capybara (2.18.0 -> 3.10.1):
Not really sure what's going on here. Looks like text in the expectation is being truncated?!?
let(:story_attributes) do
{
title: 'Edited title',
description: 'Edited location',
start_year: '2001',
start_month: 'December',
start_day: '5',
end_year: '2001',
end_month: 'October',
end_day: '10',
is_range: true,
cover_image: {
url: 'http://placehold.it/edited.png'
}
}
end
...
within 'section.story-cover' do
expect(page).to have_text 'Edited title'
expect(page).to have_text 'Edited location'
expect(page).to have_text 'December 5th - October 10th, 2001'
end
In first failed example (below) "Edited location" is being truncated.
In second example expect(page).to have_text 'Edited title Edited location' where only "Edited title\nEdited locat" is found.
Then there's this "JSON text must at least contain two octets" issue which may or may not be related but this used to pass before upgrading rspec-rails & capybara.ds
Thoughts?
RSpec Failures:
1) Story editing published edit story
Failure/Error: JSON.parse(response.body)
JSON::ParserError:
A JSON text must at least contain two octets!
# ./app/services/converter/image_service.rb:36:in `post_to_filepicker'
# ./app/services/converter/image_service.rb:18:in `convert_format'
# ./app/services/converter/image_service.rb:11:in `block in convert'
# ./app/services/converter/image_service.rb:10:in `each'
# ./app/services/converter/image_service.rb:10:in `convert'
# ./app/models/images/image.rb:5:in `convert'
# ./app/models/images/image.rb:20:in `enqueue_conversion'
# ./app/services/story/updating_service.rb:14:in `update'
# ./app/controllers/stories_controller.rb:58:in `update'
# ------------------
# --- Caused by: ---
# Capybara::ExpectationNotMet:
# expected to find text "Edited location" in "Edited title\nLyla HoegerEditedDecember 5th - October 10th, 2001Download"
# ./spec/features/stories/editing_spec.rb:86:in `block (4 levels) in <top (required)>'
2) Story editing private private story should be read after editing
Failure/Error: JSON.parse(response.body)
JSON::ParserError:
A JSON text must at least contain two octets!
# ./app/services/converter/image_service.rb:36:in `post_to_filepicker'
# ./app/services/converter/image_service.rb:18:in `convert_format'
# ./app/services/converter/image_service.rb:11:in `block in convert'
# ./app/services/converter/image_service.rb:10:in `each'
# ./app/services/converter/image_service.rb:10:in `convert'
# ./app/models/images/image.rb:5:in `convert'
# ./app/models/images/image.rb:20:in `enqueue_conversion'
# ./app/services/story/updating_service.rb:14:in `update'
# ./app/controllers/stories_controller.rb:58:in `update'
# ------------------
# --- Caused by: ---
# Capybara::ExpectationNotMet:
# expected to find text "Edited title Edited location" in "Edited title\nEdited locat"
# ./spec/features/stories/editing_spec.rb:115:in `block (4 levels) in <top (required)>'

Answer 1

One of the big changes between Capybara 2.x and 3.x was that in Capybara 3.x text is returned as closely to what is displayed as possible. This means that line feeds are now included in the returned text when they would display to the user - https://github.com/teamcapybara/capybara/blob/master/UPGRADING.md. You either need to change your expected text to "Edited title\nEdited location" at spec/features/stories/editing_spec.rb:115 or if you don't care about the linefeeds you can use the :normalize_ws option => expect(page).to have_text("Edited title Edited location", normalize_ws: true)