Dean Rasheed <dean.a.rasheed@gmail.com> writes:
> OTOH, for numeric inputs, this could easily end up needing many
> thousands of divisions and it's not hard to construct inputs that take
> minutes to compute, although this is admittedly with ridiculously
> large inputs (~10^130000), and AFAICS, the performance is OK with
> "normal" sized inputs. Should we put a limit on the size of the
> inputs?
No, but a CHECK_FOR_INTERRUPTS in the loop would be well-advised,
if there's not one already inside the called functions.
> There are apparently more efficient algorithms, but I think that
> should definitely be kept out of scope for this patch.
+1
regards, tom lane