Gregory Stark <stark(at)enterprisedb(dot)com> writes:
> Well it looks like it's trying to emulate the boolean guc behaviour but with
> an extra value.
Right, that's exactly what it's doing.
> Do we list the valid values for boolean variables at all?
Very first thing in section 18.1 of TFM. The individual variable
descriptions don't repeat that, though (and shouldn't IMHO).
regards, tom lane