Re: age(xid) on hot standby

Simon Riggs <simon(at)2ndQuadrant(dot)com> writes:
> I think we should apply the patch to return the correct SQLCODE in all
> cases, even if its supposedly not possible.

[ shrug... ] My opinion about that has not changed. Those are internal
sanity checks, and as such, ERRCODE_INTERNAL_ERROR is exactly the right
thing for them. If there are paths that can reach that code, we need to
find them and plug the holes with appropriate user-facing error checks
that say what it is the user is not supposed to do. In this example,
if we had decided that the right answer should be for age() to not be
allowed on standbys, then an error saying exactly that would be an
appropriate user-facing error. "You're not supposed to acquire a
transaction ID" is not intelligible to the average user, and giving it
another error code doesn't improve that situation.

regards, tom lane