Re: Changed SRF in targetlist handling

"David G. Johnston" <david(dot)g(dot)johnston(at)gmail(dot)com> writes:
> On Friday, June 3, 2016, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us
> <javascript:_e(%7B%7D,'cvml','tgl(at)sss(dot)pgh(dot)pa(dot)us');>> wrote:
>> Merlin Moncure <mmoncure(at)gmail(dot)com> writes:
>>> another interesting case today is:
>>> create sequence s;
>>> select generate_series(1,nextval('s')), generate_series(1,nextval('s'));

> If taking the 2.5 approach this one would fail as opposed to being
> rewritten.

Well, it'd be rewritten and then would fail at runtime because of the SRF
calls not producing the same number of rows. But even option #3 would not
be strictly bug-compatible because it would (I imagine) evaluate the
arguments of each SRF only once. The reason this case doesn't terminate
in the current implementation is that it re-evaluates the SRF arguments
each time we start a SRF over. That's just weird ...

regards, tom lane