.dt error when handling Series across daylight savings boundary

[nasrallah]

import pandas as pd
## two timepoints in UTC flanking the US/Eastern DST shift
date_strings = ['2015-11-01T05:00:00Z', '2015-11-01T07:00:00Z']
S = pd.Series(date_strings)
## Convert to US/Eastern 
S = pd.to_datetime(S).apply(lambda x: x.tz_localize('UTC').tz_convert('US/Eastern'))
S.dt.dayofweek
## AttributeError: Can only use .dt accessor with datetimelike values

## Note: the problem does not arise if the US/Eastern version of the times are on different days
## e.g. if date_strings = ['2015-11-01T03:00:00Z', '2015-11-01T07:00:00Z']

Using python 3.4 and pandas 0.17.0.
When a Series contains two time-zone aware datetimes on the 'same day' in the local time zone on either side of the Fall 2015 daylight savings time shift, calling .dt produces an attribute error.

[jorisvandenbossche]

This works in master, and this was probably fixed by #11301 (so it will be fixed in 0.17.1, to be released this weekend)

[jorisvandenbossche]

But thanks for reporting!