On Mon, Jun 11, 2018 at 10:04 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Given that this has been broken since forever, and there've been
> no complaints before now, I do not think the case for back-patching
> is strong enough to justify the problems it would cause. Just
> put it in v11 and be done.
I'm not sure I understand what problem would be created by
back-patching. It is true that anyone writing code that must work
with any version of PostgreSQL wouldn't able to count on this being
there, but the chances that anyone is writing such software using ECPG
is remote. In other words, nobody's going to add a version number
test to their ECPG code. They're just going to apply the update and
then use the new function.
I don't think this is a very important issue so I'm not prepared to
fight about it, but this looks pretty low-risk to me.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company