Jeff Davis <pgsql@j-davis.com> writes:
> BPCHAR is documented as an alias to CHARACTER, but the former can end
> up with no typmod at all whereas the latter defaults to CHARACTER(1).
> Should we document the case of using BPCHAR with no length?
Perhaps; you got specific wording in mind?
(I think that that difference is in fact documented somewhere,
but maybe not where you looked.)
regards, tom lane