Tom Lane writes:
>> Shouldn't the "public." be left out?
>
> I'm hesitant to remove the schema spec for fear we'd end up with underspecified output
> in some corner case or other.
Totally understand. How about making it an option?
Just like the --no-owner option. There are options that one may rarely use,
but that can be very usefull for certain type of specialized restores.