mehakmeet commented on PR #10834:
URL: https://github.com/apache/gravitino/pull/10834#issuecomment-4313304824

   @bharos and @mchades I understood the normalization issue clearly now. I 
have come up with a fix, please tell me if this makes sense:
   
   Wiring | Manager stores | Event sees | Hook sees | setOwner uses
   -- | -- | -- | -- | --
   OLD (pre-PR): Event → Normalize → Hook → Manager | normalized | raw | 
normalized | normalized (correct)
   Current PR (Cat/Sch/Meta): Hook → Normalize → Event → Manager | normalized | 
normalized | raw | raw (BUG)
   Current PR (Tbl/Fset/Top/Mod): Hook → Event → Normalize → Manager | 
normalized | raw | raw | raw (BUG)
   Option A (rewire): Normalize → Hook → Event → Manager | normalized | 
normalized (changed) | normalized | normalized (correct)
   
   Does Wrapping NormalizeDispatcher outermost makes sense? I think it will 
still have some gaps like the events also seeing the normalized identifier now. 
   Also I see one more issue in this, that if normalization fails due to 
invalid naming we won't be able to raise a failure event from the event 
dispatcher as it's the outermost layer. 
   But, we can have a manual normalization solution which can fill these gaps:
   
   Wiring | Manager stores | Event sees | Hook sees | setOwner uses
   -- | -- | -- | -- | --
   Option B (manual): Hook → Event → Normalize → Manager + duplicate normalize 
logic in Hook for setOwner | normalized | raw (preserves OLD) | raw | 
normalized via manual applyCapabilities (correct)


-- 
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