[Numpy-discussion] Changeset 6557
Sun Mar 8 01:12:52 CST 2009
On Sun, Mar 8, 2009 at 3:49 PM, Charles R Harris
> So it's off to look at the
> tickets and trying to fix bugs. Urrgh. Oh, and I suppose I should look into
> the argmax/argmin functions and see how they handle nans.
I think they don't at the moment: they have an implementation defined behavior.
If you're in for ref count fun, there is a bug which I did not manage
to squash when I looked at it:
More information about the Numpy-discussion