On 9/29/20 10:21 PM, Tom Lane wrote:
>
> If you ask me, the proximate cause of this bug was the decision
> to stick attrmissing info into the "constr" field of tupledescs.
> That seems pretty damfool, because there are large parts of the
> system that consider that the constr info is optional and can
> be discarded, or never built to begin with. Perhaps it's too late
> to revisit that, but we need to take a very very hard look at an
> awful lot of places if we're to stick with that design.
>
:-(
At nearly three years remove my memory is a bit hazy, but IIRC I looked
at all the places (or all I could identify) where it might have mattered
and it looked OK. But I'm not going to bet my life I caught everything,
and maybe I unintentionally provided for a future footgun, which from
your description this appears to be.
I'm not sure what the implications of changing it would be.
cheers
andrew
--
Andrew Dunstan https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services