Of course, this fix will not need to be included in any current supported release. But, would you remove the TIME WITH TIME ZONE type in a new major release?
What is the priority for PostgreSQL, standards conformance, or making a useful product without broken features?
Actually, removing it would FORCE people to fix their applications; people are already allowed to fix their own applications.
In any case this isn't a bug report - the software is working exactly as designed and, furthermore, to the letter of the specification to which it attempts to conform.
The only thing I'd consider, if I felt mis-use of this was pervasive, would be converting the plain sentence on this topic to a callout "Note" to emphasize the risk. The risk of making standard-conforming theoretically broken applications actually broken outweighs removing a marginal foot-gun.
