Be more specific regarding which exception to catch in data_utils.py - #4624

@astrofrog - I don't think in this particular case it is possible to be more specific because the input here is a list of arbitrary user functions that can fail in arbitrary ways. So for instance I might supply a function like:

def mean_inv(x):
    return np.mean(1.0 / x)

In the case where the data includes a zero then a ZeroDivisionError would be raised and not caught. The intent here was that the info attribute never fails to show something, and shows -- where the attribute is "not applicable".

Ok - in that case, I'll close this. Thanks!