potiuk opened a new pull request, #73576:
URL: https://github.com/apache/airflow/pull/73576
`handle_event_submit` only needs to add the trigger event to a task
instance's stored
`next_kwargs`. It previously decoded the whole stored payload, added the
event, and
re-encoded everything — which reconstructs every stored value in the
scheduler,
triggerer and API server, although only the worker ever needs those objects.
This merges the serialized event into the stored serialized dict instead.
Stored values
pass through untouched (the worker deserializes them as before); payloads
that are not
a dict at the top level still fail the task instance rather than resuming
it. Rows still
in the legacy `{"__type": "dict", "__var": ...}` shape keep that shape.
Test plan:
- [x] Stored kwargs carrying a nested serialized object are preserved
verbatim and the
object is never constructed during event submission
- [x] Legacy-format stored kwargs get the event added in the same format
- [x] Existing unusable-kwargs and unserializable-payload cases still fail
the task
- [x] `airflow-core/tests/unit/models/test_trigger.py` passes
##### Was generative AI tooling used to co-author this PR?
- [X] Yes — Claude Code (Opus 5.5)
Generated-by: Claude Code (Opus 5.5) following the guidelines at
https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#gen-ai-assisted-contributions
--
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]
