Hello,
>> - I do not think that updating pgbench arithmetic for managing integer
>> overflows is worth Andres Freund time. My guess is that most
>> script would not trigger client-side overflows, so the change would
>> be a no-op in practice.
>
> It might not be if you view it in isolation (although I'm not
> convinced). The problem is that it has cost beyond pgbench. Due to
> pgbench's overflow handling
Lack of?
> I can't run make check on a build that has -ftrapv, which found several
> bugs already.
Hmmm. You suggest that integer overflows do occur when running pgbench.
Indeed, this tap test case: "\set maxint debug(:minint - 1)"
Otherwise, some stat counters may overflow on very long runs? Although
overflowing a int64 takes some time...
> I'd be happy if somebody else would tackle the issue, but I don't quite
> see it happening...
I must admit that it is not very high on my may-do list. I'll review it if
it appears, though.
--
Fabien.