Re: RangeTblEntry.inh vs. RTE_SUBQUERY

On 29.02.24 19:14, Tom Lane wrote:
> Peter Eisentraut <peter(at)eisentraut(dot)org> writes:
>> In nodes/parsenodes.h, it says both
>> This *must* be false for RTEs other than RTE_RELATION entries.
>
> Well, that's true in the parser ...
>
>> and also puts it under
>> Fields valid in all RTEs:
>> which are both wrong on opposite ends of the spectrum.
>> I think it would make more sense to group inh under "Fields valid for a
>> plain relation RTE" and then explain the exception for subqueries, like
>> it is done for several other fields.
>
> Dunno. The adjacent "lateral" field is also used for only selected
> RTE kinds.

The section is

/*
* Fields valid in all RTEs:
*/
Alias *alias; /* user-written alias clause, if any */
Alias *eref; /* expanded reference names */
bool lateral; /* subquery, function, or values is
LATERAL? */
bool inh; /* inheritance requested? */
bool inFromCl; /* present in FROM clause? */
List *securityQuals; /* security barrier quals to apply, if
any */

According to my testing, lateral is used for RTE_RELATION, RTE_SUBQUERY,
RTE_FUNCTION, RTE_TABLEFUNC, RTE_VALUES, which is 5 out of 9 possible.
So I think it might be okay to relabel that section (in actuality or
mentally) as "valid in several/many/most RTEs".

But I'm not sure what reason there would be for having inh there, which
is better described as "valid for RTE_RELATION, but also borrowed by
RTE_SUBQUERY", which is pretty much exactly what is the case for relid,
relkind, etc.