Typing difference

[sklam]

The following typing difference was reported by a user:

1. divide int64 by float32 should yield float32

import numpy as np
import numba

a = np.array([1, 2, 3], dtype=np.int64)
b = np.array([4.4, 5.5, 6.6], dtype=np.float32)

def foo():
    return a / b


jit_foo = numba.njit(foo)

print(foo().dtype)
print(jit_foo().dtype)

2. invert on boolean should yield integer.

So, this means there is no ~ on bools.

import numba

def foo(a):
    return ~a

jit_foo = numba.njit(foo)

print(foo(True))
print(jit_foo(True))