Bruce Momjian <pgman@candle.pha.pa.us> writes:
> I am not either. How do you do the documentation when the function can
> be called two ways.
You don't. There is only one supported name, so that's the only one
you document.
> I guess we can give the SQL query to fix it during
> beta2 _and_ add a regression test to make sure it is fix. That sounds
> like a plan.
That sounds like massive overkill. Just apply the patch. We don't need
to institutionalize a regression test for this.
regards, tom lane