> >> In the second place, the reason most of our messages don't already
> >> contain schema names is that in the past we've judged it would be
> >> mostly clutter; and given the infrequency of complaints I see no
> >> reason to change that opinion.
>
> > I tend to disagree. We can run a poll in a wider audience.
>
> We already have a large poll: divide the number of complaints on this
> topic since 7.3 came out by the number of users ...
Since it seems like we are voting (!) let me say that fully informative
errors that include the schema would be very useful for those of us who
do use schemas to organize their tables. The generic "proper" way to
address a table in a schema (short of user path settings) is to qualify
it by its schema, so that's the unique fully descriptive name of the
table so all errors/diagnostics should reference that. Otherwise schemas
look like they are delegated to a second-class feature ("we have it so
we can check off a feature matrix, but our heart is not fully in it"). I
suspect lack of complaints is largely due to the (small) number of
people using namespaces -- the denominator should be users of the
feature, not all users...
George
