DEP: Finish deprecation in issubdtype #14619
Description
I noticed that the issubdtype function returns inconsistent output for float32, int32 and int8. I did not investigated more to see if this affects other dtypes...
Reproducing code example:
import numpy as np
np.issubdtype('float32', np.float64) # False
np.issubdtype('float32', np.double) # False
np.issubdtype('float32', np.float) # True
np.issubdtype('float32', 'float64') # True
np.issubdtype('int32', np.int64) # False
np.issubdtype('int32', 'int64') # True
np.issubdtype('int8', np.int64) # False
np.issubdtype('int8', 'int64') # True
np.issubdtype('int8', np.int32) # False
np.issubdtype('int8', 'int32') # True
np.issubdtype('int8', np.int16) # False
np.issubdtype('int8', 'int16') # TrueNumpy/Python version information:
1.16.4 3.7.3 (default, Mar 27 2019, 22:11:17) [GCC 7.3.0]