Keith Fiske <keith@omniti.com> writes:
> Just found something else rather disturbing. If you try to exclude the
> schema that the extension tables are in, their data is still output.
This is a common misconception: extensions do not live within schemas.
(An extension might own a schema, not the other way around.) So a
dump with a -n switch is never going to select an extension.
By and large, if the current behavior bothers you, ISTM it probably
means you are using these tables in a way other than what the concept of
an extension configuration table was meant for: namely, to hold
configuration data that would be referenced by the functions in that
extension, but would not normally be considered part of the user's data.
There has been some talk of trying to cater for a more general notion of
tables created by extensions, but we do not have a design or even a
clear idea of a set of requirements for that. Perhaps it would be good
if you explained what is your use-case --- why are you concerned about
being able to manage these tables as if they were regular data?
regards, tom lane
