henry3260 commented on code in PR #65587:
URL: https://github.com/apache/airflow/pull/65587#discussion_r4012860413


##########
contributing-docs/31_task_execution_architecture.rst:
##########
@@ -41,9 +41,10 @@ The two processes talk over a socket, and the Supervisor is 
the only side that e
 task JWT or talks to the *Execution API* — the user's code never sees the 
token and never touches the
 database.
 
-The same runtime can also run *in-process* (a single Python process, no fork, 
no sockets, no HTTP) for
-``dag.test()`` and local runs. The diagram below contrasts the two paths and 
marks where each Python process
-lives:
+The same runtime can also run *in-process* (a single Python process, no fork, 
no HTTP) for
+``dag.test()`` and local runs. A supervisor socket is still set up, because 
operators such as

Review Comment:
   > The figure referenced three lines below still asserts what this sentence 
now denies: the generator at 
`contributing-docs/images/diagram_task_sdk_execution_architecture.py:234` 
renders the cluster label "ONE Python process / no fork / no sockets / no 
HTTP", and `:256` labels the `InProcessSupervisorComms` node "in-memory deques, 
not sockets", which is the deque this PR removes. The 
`generate-airflow-diagrams` prek hook rebuilds the PNG but only fires on 
`^contributing-docs/images/diagram_[^/]*\.py$`, so editing those two strings is 
what gets the image regenerated. Line 44 also still says "no fork" while the 
next sentence describes an operator spawning its own child process, so 
something like "one Python process for the task itself, no HTTP" would settle 
it.
   
   Done!



-- 
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: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to