SpiffWorkflow
SpiffWorkflow copied to clipboard
Improved Error Handling and Error States
Presently in Task.py on line 749 we mark a task as complete, BEFORE we try to execute it in some way (calling the on_complete_hook). This prevents us from handling exceptions and errors in a sensible way.
This is a discussion ticket, we should brain storm a better way to handle failed executions, and assure we do not fail, and then leave ourselves in a state that prevents a rational recovery.
To be clear - we don't have a case where this is presenting a problem, failures generally are handled properly, but something here doesn't feel safe or correct, and we'd like to refactor it, and assure a safe environment.