Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Tom Lane wrote:
>> That might be a good restriction too (on top of my speculation about not
>> allowing it in postgresql.conf). Only allow it to be SET per-session,
> We don't have a way to make something unsetable in postgresql.conf right
> now, do we?
Yeah, we do --- see GUC_DISALLOW_IN_FILE. The existing variable
zero_damaged_pages has the same restrictions we're talking about here,
and for largely the same reasons: you can shoot yourself in the foot
with it.
regards, tom lane