Mem_alloc called with inconsistent type

[robertmaxton42]

~/anaconda3/envs/tensorflow/lib/python3.6/site-packages/reikna/cluda/cuda.py in __init__(self, size)
     56 
     57     def __init__(self, size):
---> 58         self._buffer = cuda.mem_alloc(size)
     59         self.size = size
     60 

ArgumentError: Python argument types in
    pycuda._driver.mem_alloc(numpy.int64)
did not match C++ signature:
    mem_alloc(unsigned long)

-- I should confess that this only showed up because of a separate bug in my own code - I was passing in an array much smaller than the Computation was asking for, so it was reallocating, I believe. But the error looked to me like a real bug I had happened to stumble on by mistake - I'm not sure when this code is meant to be called in normal operation, but I think that should be uint64 up there? - so I thought I'd report it anyway, so you knew about it.