segfault on set_index

[TomAugspurger]

I probably shouldn't have been doing this in the first place, but I'm able to consistently generate a segfault with:

import pandas as pd
pd.__version__ # '0.11.0.dev-809b238'

d = {'t1': [2, 2.5, 3], 't2': [4, 5, 6]}
df = pd.DataFrame(d)
tuples = [(0, 1), (0, 2), (1, 2)]
df['tuples'] = tuples

df.set_index(pd.MultiIndex.from_tuples(df['tuples']))

#3028 is probably relevant. I have epsilon experience with Cython so I won't really be able to help out with this one. Happy to do more debugging if you need any though.