gabotorresruiz commented on PR #43805:
URL: https://github.com/apache/superset/pull/43805#issuecomment-5623568703

   Beyond the three comments above (all fixed in 3d76c017d9), I did a full 
self-review pass of the whole export stack for the remaining instances of the 
issue classes raised across the review, and fixed a few in 40ebcc5926:
   
   - The KV distributed lock stored a tz-aware UTC expiry while the key_value 
store compares naive local (is_expired / delete_expired_entries), so on a 
non-UTC server the default non-Redis backend mis-expired the lock: the 
per-user+dashboard throttle failed open, or a killed worker held it past the 
TTL. Now naive local, matching every other writer. (This is shared infra the 
feature relies on for its throttle backstop; the Redis path was already immune.)
   - The guest success email passed the unclamped 24h ttl for a link clamped to 
one hour (latent until guest emails exist); now uses the clamped ttl so 
duration and timestamp agree.
   - export_xlsx_status reported ready when the backend was cleared after 
upload, though download_xlsx 501s that case; status now treats an unset backend 
as not-ready.
   - The throttle 202 (nothing to poll) showed a success toast; now an info 
toast.
   - Dropped a stale pre-signed-URL comment in email.py.
   
   Intentionally left, for the record: S3 AccessDenied is not mapped to 410 
(documented as needing s3:ListBucket, per our earlier thread); anonymous/Public 
sessions share lock slot 0 (documented, no identity to distinguish); 
WEBDRIVER_BASEURL_USER_FRIENDLY default is an operator requirement shared with 
scheduled reports (and the polling path uses a root-relative URL regardless). 
Tests added for each fix.


-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to