[Numpy-discussion] Do we want scalar casting to behave as it does at the moment?

Nathaniel Smith njs@pobox....
Mon Jan 7 14:16:03 CST 2013


On Mon, Jan 7, 2013 at 2:17 AM, Olivier Delalleau <shish@keba.be> wrote:
> Hehe, I didn't even know there was supposed to be a warning for arrays... Ok.
>
> But I'm not convinced that re-using the "overflow" category is a good
> idea, because to me the overflow is typically associated to the result
> of an operation (when it goes beyond the dtype's supported range),
> while here the problem is with the unsafe cast an input (even if it
> makes no difference for addition, it does for some other ufuncs).

Right, there are two operations: casting the inputs to a common type,
and then performing the addition. It's the first operation that rolls
over and would trigger a warning/error/whatever, not the second.

-n


More information about the NumPy-Discussion mailing list