Closed ccamel closed 4 years ago
@ccamel interesting concept. But im worried about two things here:
It seems to require a more explicit output mechanism for branches like we have for pipelines.
@creactiviti Yes. I see your point. You are right.
So you think of adding an explicit definition of outputs at the fork?
Something like that?
tasks:
- type: fork
label: afork
outputs:
- name: adapter1
value: ${branches[0]/adapter1}
- name: adapter2
value: ${branches[1]/adapter2}
branches:
- - name: adapter1
label: Web Service Call
type: adapter1
- - name: adapter2
label: Web Service Call
type: adapter2
- type: checkAdapterResults
name: checkAdapterResults
label: Checking Adapter Results
adapter1: ${afork.adapter1}
adapter2: ${afork.adapter2}
The PR collects the outputs of all executed tasks within a
fork
task and put them in the context. By doing this, the workflow described in #31 works fine:However, I dont't know if this behaviour is desirable, or if it is wise to do so. @creactiviti I let you decide what to do. :wink: