Hello Robert,
>>> And this one generates a core dump:
>>> \set cid debug(-9223372036854775808 / -1)
>>> Floating point exception: 8 (core dumped)
>
> That does not seem acceptable to me. I don't want pgbench to die a
> horrible death if a floating point exception occurs any more than I want
> the server to do the same. I want the error to be properly caught and
> reported.
Please note that this issue/bug/feature/whatever already exists just for
these two values (1/2**128 probability), it is not related to this patch
about functions.
sh> pgbench --version pgbench (PostgreSQL) 9.5.0
sh> cat div.sql \set i -9223372036854775807 \set i :i - 1 \set i :i / -1
sh> pgbench -f div.sql starting vacuum...end. Floating point exception (core dumped)
I do not think that it is really worth fixing, but I will not prevent
anyone to fix it.
--
Fabien.
