Series.resample leads to RuntimeError: maximum recursion depth exceeded while calling a Python object

[cancan101]

rng = pd.date_range('1/1/2012', periods=4, freq=pd.offsets.YearEnd())
s = pd.Series(np.random.randn(4), index=rng)
s.resample(rule=pd.offsets.Week(weekday=0), fill_method="pad", closed="left", kind="period")

leads to:

...
/usr/local/lib/python2.7/dist-packages/pandas/tseries/period.pyc in _from_arraylike(cls, data, freq, tz)
    652                                       'inferred from first element'))
    653 
--> 654                 if np.issubdtype(data.dtype, np.datetime64):
    655                     data = dt64arr_to_periodarr(data, freq, tz)
    656                 elif data.dtype == np.int64:

/usr/local/lib/python2.7/dist-packages/numpy/core/numerictypes.pyc in issubdtype(arg1, arg2)
    757 
    758     """
--> 759     if issubclass_(arg2, generic):
    760         return issubclass(dtype(arg1).type, arg2)
    761     mro = dtype(arg2).type.mro()

/usr/local/lib/python2.7/dist-packages/numpy/core/numerictypes.pyc in issubclass_(arg1, arg2)
    697     """
    698     try:
--> 699         return issubclass(arg1, arg2)
    700     except TypeError:
    701         return False

Setting kind="period" seems to trigger the issue.

[cpcloud]

already have an issue for this #3612