Pluggable result cache for data being returned from jobs

The Gearman Job Server keeps an internal hash table based on unique job ID. This feature would enable that hash to have pluggable storage to keep output from jobs. Each module could store and manage that cache in any way they like (memory, file, or DB storage, implement TTLs, filters, ...). When a new job request comes in for a unique key that is in the hash table, the job can optionally be bypassed and the result can simple be returned from the cache instead. This gives memcached like ability to gearman, but in a pluggable way.