None not always considered missing

[nspies]

I expected pandas to know that None should be ignored as a missing value in the following simple example:

import pandas
import numpy

# works fine
df = pandas.DataFrame({"col1":[2,5.0,123,numpy.nan], "col2":[1,2,3,4]})
print df.sub(3)

# fails -- tries to do the subtraction with None
df = pandas.DataFrame({"col1":[2,5.0,123,None], "col2":[1,2,3,4]})
print df.sub(3)

[wesm]

I made the requisite changes here in master, will be part of 0.7.3 release