Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
> I have the attached patch that would make character_octet_length the product
> of character_octet_length and the maximum octet length of a single character
> in the selected server encoding. So for UTF-8, this would be factor 4. This
> doesn't exactly correspond to the behavior that you expect, but I think it's
> more correct overall anyway.
+1, but that new query isn't very schema-safe ... I think it needs a few
"pg_catalog." qualifications.
regards, tom lane