David Fetter <david@fetter.org> writes:
> On Fri, Mar 28, 2014 at 03:50:10PM +0000, Tom Lane wrote:
>> Fix EquivalenceClass processing for nested append relations.
> If this is a bug fix, shouldn't it be back-patched?
It's a planner improvement; it's not a bug fix in the sense that previous
plans were wrong. We tend to err on the side of not back-patching such
changes ... some people complain when minor updates cause plan choices to
change under them.
regards, tom lane