Christopher Kings-Lynne <chriskl@familyhealth.com.au> writes:
> Resubmission of patch (for 7.4).
While I agree with the objection that we shouldn't break the strings
freeze for this, I do think it might be a good idea to apply the change
that prints CHECK constraints using pg_get_constraintdef() rather than
simply looking at pg_constraint.consrc. This is arguably a bug fix,
not a pretty printing improvement, because consrc doesn't track column
renamings (and probably other changes).
Doing that would require making this one string change:
! printfPQExpBuffer(&buf, _(" \"%s\" CHECK %s"),
becomes
! printfPQExpBuffer(&buf, _(" \"%s\" %s"),
but ISTM the latter string doesn't really need any translation and so
it wouldn't be much of a problem.
Comments?
regards, tom lane