Question

I want to iterate over a JSON-API every 10th second, and do a second HTTP-request using the same connection (keepalive) if a certain key was found in the JSON-data. If I do not place EM.stop in my code, the program stops to wait after finishing processing in req1.callback.

If I put EM.stop inside req2.callback it works and is iterating as expected.

But if the JSON-document did not include the key foobar, the program stops to wait after finishing processing in req1.callback.

If I add EM.stop in the last line inside req1.callback, req2.callback is aborted if the JSON-document had the key foobar.

How should I properly place EM.stop to make it iterate if the JSON-document had what I wanted or not?

require 'eventmachine'
require 'em-http'

loop do    
  EM.run do
    c = EM::HttpRequest.new 'http://api.example.com/'

    req1 = c.get :keepalive => true
    req1.callback do
      document = JSON.parse req1.response
      if document.has_key? foobar   
        req2 = c.get :path => '/data/'
        req2.callback do
          puts [:success, 2, req2]
          puts "\n\n\n"
          EM.stop
        end
      end
    end
  end

  sleep 10
end

Answer 1

If you want to use a timer, you should use the actual timer support from EM: http://eventmachine.rubyforge.org/EventMachine.html#M000467

For example:

require 'eventmachine'
require 'em-http'

EM.run do
  c = EM::HttpRequest.new 'http://google.com/'
  EM.add_periodic_timer(10) do
    # Your logic to be run every 10 seconds goes here!
  end
end

This way, you keep EventMachine running the whole time, instead of having to start/stop it every 10 seconds.