Amit Kapila <amit.kapila16@gmail.com> writes:
> Fair enough. Do you mind being consistent in this regard for logical
> replication-related code?
As long as not too many individual changes are involved, sure.
But consistency of this sort doesn't seem worth creating a lot
of back-patching land mines, IMO.
> BTW, is there a reason we prefer to write in
> one or another way (with or without appending schema_name)?
In places where we can trust the search_path to be just pg_catalog,
there's no real strong reason to write a schema name, I think.
regards, tom lane