Peter Eisentraut <peter.eisentraut@2ndquadrant.com> writes:
> On 7/26/16 6:14 PM, Vik Fearing wrote:
>> As mentioned elsewhere in the thread, you can just do WHERE true to get
>> around it, so why on Earth have it PGC_SUSET?
> I'm not sure whether it's supposed to guard against typos and possibly
> buggy SQL string concatenation in application code. So it would help
> against accidental mistakes, whereas putting a WHERE TRUE in there would
> be an intentional override.
Maybe I misunderstood Vik's point; I thought he was complaining that
it's silly to make this SUSET rather than USERSET. I tend to agree.
We have a rough consensus that GUCs that change query semantics are
bad, but if it simply throws an error (or not) then it's not likely
to cause any surprising application behaviors.
regards, tom lane