Vamsi-klu commented on issue #72902:
URL: https://github.com/apache/airflow/issues/72902#issuecomment-5638724649
Tested my four items on this RC. Airflow 3.3.1, Python 3.12, wheels from
PyPI.
databricks 7.20.0rc1 for #71752 and #70088. Start/Stop warehouse operators
behave as advertised. Empty warehouse_id raises, already RUNNING/STOPPED is a
no-op, sync wait polls through, and deferrable=True hands off to
DatabricksWarehouseStateTrigger. execute_complete is fine on the target state
and errors on deleted/timeout. No live workspace here, so this was mocked
against the installed hook.
google 22.5.0rc1 for #72132. The stream path now honors gzip and mime_type.
use_stream=True sets blob.content_type before opening the writer, wraps
GzipFile when gzip=True, and opens with ignore_flush=True so BlobWriter.flush
does not explode. Content-Encoding stays unset. Default is still
use_stream=False.
http 6.1.0rc1 for #69748. deferrable POST/PATCH logs one warning about
Triggerer-restart replay. The RFC idempotent methods stay quiet. Construction
and the trigger itself do not warn, and warn_on_non_idempotent=False shuts it
off. Paginated POST still warns once. One small leftover: the warning points at
the stable deferrable page, which 404s until 6.1.0 docs go out. The staged page
is already there.
microsoft-azure 15.1.0rc1 for #72130. ADF and Synapse async hooks parse
extras with json.loads(conn.extra) and never touch extra_dejson. Forced
extra_dejson to raise the AsyncToSync error and get_async_conn still succeeded.
refresh_conn closes the cached client. Empty extra becomes {}.
All four look good.
--
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]
