Marcus Gartner <m(dot)a(dot)gartner(at)gmail(dot)com> writes:
> subpath(ltree,offset,len) now correctly errors when given an offset less
> than -n where n is the number of labels in the given ltree. A duplicate
> block of code has been removed which allowed an offset as low as -2n.
Nice catch ... as far as I can see from the git history, that mistake
has been there since day one.
I'm inclined to put this in master only, without a back-patch.
It's not fixing any case that should work, and it seems possible
that somebody out there is depending on the bogus behavior.
regards, tom lane