Instance Recovery: Script Task handling

[jjoderis]

In its current form, the process instance recovery feature tries to resume execution of process elements that had a token when the instance was interrupted if possible. This works for User-Tasks, Call-Activities and Sub-Processes.
In case of Script-Tasks there is currently no way to determine automatically if the logic of the script task is idempotent and could be repeated, if it could be skipped or if more involved error handling is necessary so the task is put into an error state.

What should happen in the MS is that the user should be informed about a Script-Task and should be given options what to do (skip, restart, manual handling) and the information to determine what the best course of action would be.
The information that will be necessary at a minimum are the variables that the Script-Task might have already written into the global instance state (intermediate-variable changes should be reset when the task is restarted) and http requests that were already made.