On Sat, Aug 1, 2015 at 8:39 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> I propose to back-patch this into 9.5, but not further; it's not an
> important enough issue to justify changing SQLSTATE behavior in stable
> branches.
+1. As I've said in the past, I think that making it possible to
determine mechanically that a "can't happen" error has in fact
occurred is of huge value to the project.
--
Peter Geoghegan