Looking for a different flavor of on_start() behavior

[markcampanelli-wf]

Description of issue / feature request

My task workflow requires something like the on_start() in a super-TaskSet, but the requirement is that the on_start() function execute every time a sub-TaskSet is interrupted and returns control to the super-TaskSet. From what I can tell, the on_start() is only ever called once.

The use case is opening a document in the super-TaskSet and then, in the sub-TaskSet, manipulating and eventually closing/leaving the document (via an interrupt) and returning to the super-TaskSet where the document must then be re-opened before returning to the sub-TaskSet.

Perhaps I am missing an existing pattern for achieving this workflow?

Expected behavior

N/A

Actual behavior

A/A

Environment settings (for bug reports)

• OS: Mac
• Python version: 3.6.3
• Locust version: 1.8.1

Steps to reproduce (for bug reports)

N/A

[heyman]

Could you maybe do it in TaskSet.execute_next_task() (https://github.com/locustio/locust/blob/master/locust/core.py#L295-L297) in the super-TaskSet?

Something like:

class SuperTaskSet(TaskSet):
    def execute_next_task(self):
        # insert code to be executed before every task here
        
        super(SuperTaskSet, self).execute_next_task()

[markcampanelli-wf]

Thanks @heyman. I'll give this a try today and report back here.

[markcampanelli-wf]

@heyman This solution works and seems reasonable. Thanks.