I am getting am empty document back from phantomjs. I am trying to use Capybara and Poltergeist for setting up the phantomjs driver for Capybara.
I created a module as follows and included it in the file that uses needs to connect.
require 'capybara/poltergeist'
module Parser
module JSParser
include Capybara
# Create a new PhantomJS session in Capybara
def new_session
# Register PhantomJS (aka poltergeist) as the driver to use
Capybara.register_driver :poltergeist do |app|
Capybara::Poltergeist::Driver.new(app, :debug => true)
end
# Use XPath as the default selector for the find method
Capybara.default_selector = :xpath
Capybara.javascript_driver = :poltergeist
Capybara.current_driver = :poltergeist
# Start up a new thread
@session = Capybara::Session.new(:poltergeist)
# Report using a particular user agent
@session.driver.headers = { 'User-Agent' => 'Mozilla/5.0 (Macintosh; Intel Mac OS X)' }
# Return the driver's session
@session
end
# Returns the current session's page
def html
@session.html
end
end
end
Then, loading the page as follows:
class Loader
include Parser::JSParser
def load_page
new_session
visit "http://www.smashingmagazine.com"
#let phantomjs take its time
sleep 5
puts "html=#{html}"
end
end
Then, finally, calling the load_page
Loader.new.load_page
Here is the debug response from poltergeist
poltergeist [1364758785355] state default -> loading
{"response"=>true}
{"name"=>"visit", "args"=>["http://www.smashingmagazine.com"]}
poltergeist [1364758794574] state loading -> default
{"response"=>{"status"=>"success"}}
{"name"=>"body", "args"=>[]}
{"response"=>"<html><head></head><body></body></html>"}
As you can see, the response is just a blank document with only the html, head and body tags but nothing in the body tag.
What wrong am I doing? Observing network traffic, I am getting the full response back from the host (smashingmagazine.com in this case). Its after the response comes back that I don't know what is happening. Sometimes phantomjs is also crashing and on other occasions, it goes through with the empty body. Here is the last line that is printed on STDERR when phantomjs crashes
PhantomJS client died while processing {"name":"visit","args":["http://www.smashingmagazine.com"]}
'--ssl-protocol=TLSv1'argument solved a similar empty response issue for me. – Asbestos