Changes between Initial Version and Version 1 of Ticket #34200, comment 11


Ignore:
Timestamp:
Dec 6, 2022, 1:15:51 AM (2 years ago)
Author:
Mike Crute

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #34200, comment 11

    initial v1  
    11Replying to [comment:7 Florian Apolloner]:
    2 > I think this makes sense and is hard to achieve otherwise short of overriding the backend. I am wondering if explicitly supporting role makes sense though. Would a more generic approach like `init_command` in mysql help? Then again we never had the need for this in postgresql so I rather explicitly support `role` (or `assume_role` maybe) -- this way we can easily shuffle it around in the backend code if needed.
     2> I am wondering if explicitly supporting role makes sense though. Would a more generic approach like `init_command` in mysql help? Then again we never had the need for this in postgresql so I rather explicitly support `role` (or `assume_role` maybe) -- this way we can easily shuffle it around in the backend code if needed.
    33
    44I don't think this is a general concept, it's pretty specific to the postgres security model. I'm happy to call it either `role` or `assume_role` if you have a preference (I do not). Given that it's pretty postgres specific that would mean it should stay in `OPTIONS` right?
Back to Top