Neil Conway <nconway(at)klamath(dot)dyndns(dot)org> writes:
>> This is better than what we had, but still seems confusing
> Ok, I made this part of the documentation more clear.
BTW, the text still fails to make it clear that the statements will in
fact reset FOUND to false when they process zero rows (or whatever).
One could easily read this text as saying that FOUND will be the logical
OR of whether all the statements-so-far found a row.
It might also be good to spell out that FOUND is local in each plpgsql
function, not a global variable.
regards, tom lane