Re: ruleutils vs. empty targetlists

Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com> writes:
> On 13 December 2013 15:07, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> How about as-is in the back branches, and throw the new errors only
>> in HEAD?

> Seems reasonable.

After further experimentation I've concluded that maybe we'd better
not back-patch this change. The reason for my change of heart is
that in 8.4, the patch made plpgsql stop complaining that "return ;"
was missing an expression. While we could possibly fix that, or
just decide not to patch 8.4, it occurs to me that there might be
applications out there that are expecting "SELECT ;" to fail, too.
So the risk of undesirable behavior changes seems a little larger
than I'd previously believed, and I'm feeling that fixing this
corner case in the back branches is not worth that risk.

regards, tom lane