Not compatible with resque-status

Resque-loner does not work when also using the resque-status plugin since the latter adds a job_id to the payload and so that confuses resque-loner into thinking that the job does not already exist and thus allows. I am not sure there’s much to do about this but I’m just opening this ticket as a reference for others.

For now, I’m trying to work around the issue by monkey patching the redis_key function to remove the job_id from the payload before calculating the md5 hash.

Here’s the code:

# Monkey patch resque-loner so that it plays ball with resque-status.
module Resque
  module Plugins
    module UniqueJob
      module ClassMethods
        def redis_key(payload)
          payload = decode(encode(payload)) # This is the cycle the data goes when being enqueued/dequeued
          job  = payload[:class] || payload['class']
          args = (payload[:args]  || payload['args'])

          # This is a quick hack so that this library works with resque-status.
          # Resque-status adds a job_id as the first element of args, and this
          # messes up how resque-loner determines if a job is already queued or
          # not. By removing the job_id, everything should be A-OK. Only workers
          # that include resque-loner should be affected. Small caveat, when calling
          # Resque.enqueued? that comes with resque-loner, there’s no job_id to
          # remove so we must do nothing in that case.
          if respond_to?(:create)
            args.shift unless caller.index { |e| e =~ /enqueued\?/ }
          end

          args.map! do |arg|
            arg.is_a?(Hash) ? arg.sort : arg
          end

          digest = Digest::MD5.hexdigest(encode(class: job, args: args))
          digest
        end
      end
    end
  end
end

[jchatel]

Still the case in end of 2015 unfortunately.