[PR] Fix get_leaves calculation for teardown in nested group [airflow]

[via GitHub]

dstandish opened a new pull request, #36456:
URL: https://github.com/apache/airflow/pull/36456

   When arrowing `group` >> `task`, the "leaves" of `group` are connected to 
`task`. When calculating leaves in the group, teardown tasks are ignored, and 
we recurse upstream to find non-teardowns.
   
   What was happening, and what this fixes, is you might recurse to a work task 
that already has another non-teardown downstream in the group.  In that case 
you should ignore the work task (because it already has a non-teardown 
descendent).
   
   ### Before:
   
![image](https://github.com/apache/airflow/assets/15932138/15da394b-c619-4dc6-a430-847dbfb2781f)
   
   
   ### After:
   
![image](https://github.com/apache/airflow/assets/15932138/7d7021f6-bc27-4063-8598-95f94b005322)
   
   Code:
   ```python
   with DAG(dag_id="s_t_dag") as dag:
   
       @task
       def test_task():
           print("Hello world!")
   
       @task_group
       def inner():
           inner_start = EmptyOperator(task_id="start")
           inner_end = EmptyOperator(task_id="end")
   
           test_task_r = test_task.override(task_id="work")()
           inner_start >> test_task_r >> 
inner_end.as_teardown(setups=inner_start)
   
       @task_group
       def outer():
           outer_work = EmptyOperator(task_id="work")
           inner_group = inner()
           inner_group >> outer_work
   
       dag_start = EmptyOperator(task_id="dag_start")
       dag_end = EmptyOperator(task_id="dag_end")
       dag_start >> outer() >> dag_end
   ```
   
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscr...@airflow.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org