Setting page/respone.body in Capybara used to work in Webrat - ruby-on-rails

Question

I am migrating to Capybara.
One of the problems I have is migrating the pdf step.
This step sets page.body to a parsed pdf.
That way I can use the default cucumber steps.
When 'I follow the PDF link "$label"' do |label|
click_link(label)
page.body = PDF::Inspector::Text.analyze(page.body).strings.join(" ")
end
Ex.
When I follow the PDF link "Catalogue"
Then I should see "Cheap products"
The error I get is this one:
undefined method `body=' for #<Capybara::`enter code here`Document> (NoMethodError)

Answer 1

On top, make sure you set :js => true like this:
scenario 'do something', :js => true do
str = PDF::Inspector::Text.analyze(page.body).strings.join(" ") # or whatever string you want
# then just use javascript to edit or add the body
page.evaluate_script("document.write(#{str});")
end
Now this is dependent on the driver, but it's one solution...

Answer 2

There is no setter for body defined in the source in capybara, so you cannot set it externally like that. Try this (untested):
page.instance_variable_set(:#body, PDF::Inspector::Text.analyze(page.body).strings.join(" "))

Answer 3

This worked for me:
Then /^I should be served the document as a PDF$/ do
page.response_headers['Content-Type'].should == "application/pdf"
pdf = PDF::Inspector::Text.analyze(page.source).strings.join(" ")
page.driver.response.instance_variable_set('#body', pdf)
end
Then /^I should see the document details$/ do
page.should have_content("#{#document.customer.name}")
page.should have_content("#{#document.resources.first.resource.name}")
page.should have_content("Document opened at #{#document.created_at.strftime("%e-%b-%4Y %r")}")
end
Note that I'm serving my PDF inline
pdf = DocumentPdf.new(#document)
send_data pdf.render, :filename => "document_#{#document.created_at.strftime("%Y-%m-%d")}",
:type => "application/pdf",
:disposition => "inline"