nolan(at)celery(dot)tssi(dot)com writes:
> Several weeks ago I posted a note about a plpgsql function that fails
> if an index that existed when it was created is subsequently dropped.
No surprise...
> Would a cross-query plan cache for SQL functions create similar problems?
Yeah. We need to think about a mechanism for invalidating plans before
we go too much further down the road of caching plans.
regards, tom lane