liujiayi771 opened a new pull request, #12027:
URL: https://github.com/apache/gluten/pull/12027

   ## Summary
   
   When `gluten.enabled=false` (e.g. the vanilla Spark baseline run inside 
`runQueryAndCompare`), `GlutenQueryExecutionListener` still posted a 
`GlutenPlanFallbackEvent`. Because no Gluten rules ran, all nodes appeared as 
fallback with the generic reason *"Gluten does not touch it or does not support 
it"*, polluting the event stream.
   
   This was an oversight in https://github.com/apache/gluten/pull/11853, which 
introduced the fallback event posting but did not account for the case where 
Gluten is disabled entirely via session config. When Gluten is off, every node 
is trivially "fallback" — posting an event is misleading and noisy.
   
   **Fix:** Early-return before `collectQueryExecutionFallbackSummary` when the 
session's `gluten.enabled` config is `false`. The config is read directly from 
`qe.sparkSession.sessionState.conf` via `getConfString` to avoid thread-local 
`SQLConf` timing issues in the async listener thread.
   
   ## Test plan
   
   - [x] Added regression test `FallbackSuite`: verifies no 
`GlutenPlanFallbackEvent` is emitted when a query runs with 
`gluten.enabled=false`


-- 
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