Job Pre/Post processing flexibility

[aldbr]

• Currently, a VO can define 1 JobExecutionCoordinator, which happens just after the execution of the payload, before the upload of the result.
• A given VO might want to deal with output data in a different way, so the upload of the result should probably move inside the base JobExecutionCoordinator
• But a given VO might want to use specific operations from the base one, and redefine an operation in-between.

Currently, the system is not flexible enough to support that kind of use cases and should be revised.