python 2.7 range accepts numpy float16, float32, complex when it should raise TypeError

In python 2.7, range accepts numpy.float16 and np.float32, np.float128, complex64 and complex128 as inputs, which it really shouldn’t.

Reproducing code example:

>>> import numpy as np
>>> range(np.float16(4))
[0, 1, 2, 3]
>>> range(np.float32(4))
[0, 1, 2, 3]
>>> range(np.complex64(4.))
/home/mark/miniconda3/envs/np2/bin/ipython:1: ComplexWarning: Casting complex values to real discards the imaginary part
  #!/home/mark/miniconda3/envs/np2/bin/python
[0, 1, 2, 3]
>>> range(np.complex128(4.))
/home/mark/miniconda3/envs/np2/bin/ipython:1: ComplexWarning: Casting complex values to real discards the imaginary part
  #!/home/mark/miniconda3/envs/np2/bin/python
[0, 1, 2, 3]
>>> range(np.float128(1))
[0]

Success??? But these fail (as expected)

>>> import numpy as np
>>> range(np.float(4))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: range() integer end argument expected, got float.
>>> import numpy as np
>>> range(np.float64(4))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: range() integer end argument expected, got numpy.float64.

Not sure if this is a python bug, or numpy bug.

In python3:

>>> import numpy as np
>>> range(np.float16(4))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'numpy.float16' object cannot be interpreted as an integer
>>> range(np.float32(4))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'numpy.float32' object cannot be interpreted as an integer
>>> range(np.complex64(4.))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'numpy.complex64' object cannot be interpreted as an integer
>>> range(np.complex128(4.))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'numpy.complex128' object cannot be interpreted as an integer
>>> range(np.float(4))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'float' object cannot be interpreted as an integer
>>> range(np.float64(4))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'numpy.float64' object cannot be interpreted as an integer
>>> range(np.float128(1))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'numpy.float128' object cannot be interpreted as an integer

Numpy/Python version information:

Python 2.7

>>> import sys, numpy; print(numpy.__version__, sys.version)
('1.16.0.dev0+b99c73d', '2.7.15 |Anaconda, Inc.| (default, Sep 27 2018, 20:19:23) \n[GCC 4.8.2 20140120 (Red Hat 4.8.2-15)]')

or

>>> import sys, numpy; print(numpy.__version__, sys.version)
('1.15.2', '2.7.15 |Anaconda, Inc.| (default, Sep 27 2018, 20:19:23) \n[GCC 4.8.2 20140120 (Red Hat 4.8.2-15)]')

Python 3.6

>>> import sys, numpy; print(numpy.__version__, sys.version)
1.15.2 3.6.6 | packaged by conda-forge | (default, Jul 26 2018, 09:53:17) 
[GCC 4.8.2 20140120 (Red Hat 4.8.2-15)]

[seberg]

Well, python fixed it already, its part of python 3 😃.

[hmaarrfk]

It isn’t really likely that anybody outside the scientific community will want to fix this though.