Andrzej Mazurkiewicz <andrzej@mazurkiewicz.org> writes:
> After reading please comment if there are more objections for changing the
> depedency type for trigger to constraint dependency from the
> DEPENDENCY_INTERNAL to DEPENDENCY_AUTO.
I'm not sure which part of "no" you didn't understand, but we're not
doing that.
> lipa=> DROP TRIGGER RI_ConstraintTrigger_c_19322 ON master;
> ERROR: trigger "ri_constrainttrigger_c_19322" for table "master" does not
> exist
This has to do with case-folding and lack of double quotes, not anything
more subtle than that. A correct test would've given results like this:
regression=# drop trigger "RI_ConstraintTrigger_a_43528" on master;
ERROR: cannot drop trigger RI_ConstraintTrigger_a_43528 on table master because constraint fk0_detail on table detail
requiresit
HINT: You can drop constraint fk0_detail on table detail instead.
which is the behavior we need.
regards, tom lane