"Miguel Ferreira" <miguelmbferreira(at)gmail(dot)com> writes:
> I propose extending the syntax of the ALTER TABLE ADD CONSTRAINT
> statement (and potentially CREATE TABLE) to allow for the
> specification of a custom error message for each CHECK constraint.
Why don't you just choose better names for your constraints?
I'd argue that the proposed change might actually be a net loss
for usability, if it entirely obscures the fact that what happened
was a check-constraint violation.
It's also not very clear why we'd stop with check constraints,
if the desire is to get rid of database-produced error messages
in favor of something that somebody likes better.
regards, tom lane