On 24.09.24 02:30, David Rowley wrote:
> I understand the last discussion went down that route too. For me, it
> seems strange that adding some global variable is seen as cleaner than
> storing the property in the same location as all the other index
> properties.
It's arguably not actually a property of the index, it's a property of
the user's session. Like, kind of, the search path is a session
property, not a property of a schema.
> How would you ensure no cached plans are still using the index after
> changing the GUC?
Something for the patch author to figure out. ;-)