Rod Taylor <rbt@zort.ca> writes:
> The issue with finding and removing foreign key constraints is no longer
> an issue, so please apply the attached.
I do not like adding a system-table index just for the convenience of
this one not-very-time-critical function. Please rewrite the patch
without that.
BTW, if you are only reading the table and not writing it, the correct
lock type to get is AccessShareLock, not RowExclusiveLock.
regards, tom lane