Pavel Stehule <stehule(at)kix(dot)fsv(dot)cvut(dot)cz> writes:
> I wont to prohibit synonyms in exception (every exception has unique
> sqlstate).
I don't think that's a particularly good idea --- maybe if SQL had been
designed according to your worldview, it'd be like that, but it isn't
and you can't retroactively force it to be. The SQLSTATEs are
deliberately designed to be fairly coarse, not unique. I believe the
design intention is to distinguish between two cases when it's likely
that client application code would do something different in the two
cases. Not to be "unique for uniqueness' sake".
regards, tom lane