For the task cache to be invalidated when the output handling functions/templates change
What would be the benefit? Does the change make something easier to use?
Primarily when debugging a workflow with faulty task output handling. Currently, if you make changes to the output handling of a shell-command you need to remember to delete the corresponding cache directory, otherwise the outputs will be read from the previous run.
If the output spec was included in the checksum for the task then Pydra would create a new cache directory and you wouldn't run into this trap
If it is a new feature, what is the benefit?
Make it simpler for inexperienced (and experienced) users to debug task output handling within the context of a workflow
What would you like changed/added and why?
For the task cache to be invalidated when the output handling functions/templates change
What would be the benefit? Does the change make something easier to use?
Primarily when debugging a workflow with faulty task output handling. Currently, if you make changes to the output handling of a shell-command you need to remember to delete the corresponding cache directory, otherwise the outputs will be read from the previous run.
If the output spec was included in the checksum for the task then Pydra would create a new cache directory and you wouldn't run into this trap
If it is a new feature, what is the benefit?
Make it simpler for inexperienced (and experienced) users to debug task output handling within the context of a workflow