functions with ndarray parameters of type npy_bool always raise a ValueError

[telamonian]

Here's a potential bug that I came across when answering a question on Stack Overflow.

Given a boolFunc.pyx file that looks like this:

import numpy as np
cimport numpy as np

def boolFunc(np.ndarray[np.npy_bool, ndim=1] npdata):
    return npdata

if you try to import and use it in python code:

from boolFunc import boolFunc
import numpy as np

# generate random boolArr
arr = np.random.randint(0,2, size=3, dtype=int)
boolArr = np.array(arr, dtype=bool)

# test cython function
boolFunc(boolArr)

you always get a particular ValueError:

ValueError: Does not understand character buffer dtype format string ('?')

This seems like it should work as is. In fact, this does work for other numpy types (eg npy_int, npy_double, etc). On the other hand, adding cast=True to the signature of boolFunc:

def boolFunc(np.ndarray[np.npy_bool, ndim=1, cast=True] npdata):

fixes the error, but casting shouldn't be necessary here since we're explicitly passing in a bool type ndarray.