Improved README.md to explain the steps needed to start the 'resque-web' server interface properly with the 'resque-retry' features. The problems as I identified them essentially were: A. The README proposed a rack configuration, but didn't include an example command for a user to start the server. Solution: Added a simple example on how to start the server with bundler by loading the proposed rack configuration. B. The README didn't include the alternate option of starting the resque-web interface using the built-in 'resque-web' command. Solution: Included a simple 3 step explanation on how to use the resque-web command by sending it a configuration file as a parameter
…led. @retry_attempt depends on perform being called, but perform will be called after forking and set in the child job process. If there is a crash that prevents the job from notifying the failure hooks itself, then a DirtyExit exception will be given to the on_failure hooks from a worker processes, where @retry_attempt will not have been set. The @on_failure_retry_hook_already_called variable also needs to be ignored, because this is set to false in the before_perform hook.
Preserve backwards compatibility for gem-users that have overridden the "args_for_retry" method. Also, ensure that the tests for "args_for_retry", "retry_args" and "retry_args_for_exception" are not overcomplicated, all we care about is that they are hit.
…pecially useful for situations where you retry a job multiple times and the backoff period for a given run is likely to be longer than the default expire_time that was set