ExperimentalSubscription.rb

# This is how Queue#subscribe will work in the next version of Bunny (v0.5.4).
# It's in the 'experimental' branch now.
#
# In order for subscription to work smoothly you need to use a combination of
# Queue#subscribe, Queue#unsubscribe, Queue#ack and Client#qos prefetch. The
# prefetch will allow a controlled number of messages to be released to a
# consumer, which means that Queue#unsubscribe and other methods can be run
# without errors due to out of sequence messages.

#
# Example 1 - Using :message_max where no. of messages > message_max
#

# Start Bunny instance
b = Bunny.new(:logging => true)
b.start

# Create queue
q = b.queue('subtest1')

# Make sure queue is empty
q.purge

# Set prefetch window (no. of messages that server will send at a time).
# Default is 1. RabbitMQ v1.6.0 only accepts :prefetch_count option. The
# other options have not been implemented yet. 
b.qos()

# Publish some messages
10.times {q.publish('Here be another test message')}

# Subscribe to the queue. Must specify :ack => true otherwise
# prefetch will not work.
q.subscribe(:message_max => 5, :ack => true) {|msg| puts msg[:payload]} 

# Check the queue status
s = q.status
puts "Queue: #{q.name}\nMessages: #{s[:message_count]}\nConsumers: #{s[:consumer_count]}"

#
# Example 2 - Using :timeout
#
# Copy set up stuff from Example 1. Prefetch is not needed.

# Subscribe to the queue
q.subscribe(:timeout => 10) {|msg| puts msg[:payload]} 

# Check the queue status
s = q.status
puts "Queue: #{q.name}\nMessages: #{s[:message_count]}\nConsumers: #{s[:consumer_count]}"

#
# Example 3 - Only process the first message from the queue. N.B. More easily
#             done with :message_max option
#
# Copy set up stuff from Example 1. Prefetch is needed.

# Subscribe to the queue. Prefetch is needed
q.subscribe(:consumer_tag => 'my_consumer', :ack => true) do |msg|
  # Get first message
  puts msg[:payload]
  # Unsubscribe from queue
  q.unsubscribe()
  # Acknowledge the message
  q.ack()
  # Break out of the loop
  break
end

# Check the queue status
s = q.status
puts "Queue: #{q.name}\nMessages: #{s[:message_count]}\nConsumers: #{s[:consumer_count]}"

#
# Example 4 - Process all messages from queue then stop when queue empty
#

b = Bunny.new
b.start()
b.qos()

q = b.queue()
10.times { q.publish('Hey Ho') }

cnt = 0

q.subscribe(:ack => true) do |msg|
  cnt += 1
  msg_cnt = q.message_count()
  puts '************************'
  puts "Message: #{cnt} - #{msg[:payload]}"
  puts "#{msg_cnt} message(s) left in the queue"
  puts '************************'
	
  if msg_cnt == 0
    q.unsubscribe()
    q.ack()
    break
  end
	
end

#
# Example 5 - Display all of the information about a consumed message
#

b = Bunny.new()
b.start()
b.qos()

q = b.queue()
q.publish('Another message served up by RabbitMQ')

q.subscribe(:message_max => 1, :ack => true) do |msg|

  puts
  puts "Message Information"
  puts '-------------------'
  puts "Payload - #{msg[:payload]}"
  puts
  puts "Header - "
  msg[:header].instance_variables.each do |var|
    vname = var.to_s.sub(/@/, '')
    val = eval("msg[:header]." + vname)
    if val.is_a?(Hash)
      puts "  #{vname.capitalize} -"
      val.each_pair do |k,v|
        puts "    #{k.capitalize}: #{v}"
      end
    else
      puts "  #{vname.capitalize}: #{val.to_s}"
    end		
  end
  puts
  puts "Delivery details -"
  msg[:delivery_details].each_pair do |k,v|
    puts "  #{k.capitalize}: #{v}"
  end
  puts
	
end