Looplift fail on if-branch after for-loop in py3.8

[sklam]

Reporting a bug

• I am using the latest released version (0.47) of Numba (most recent is visible in
  the change log (https://github.com/numba/numba/blob/master/CHANGE_LOG).
• I have included below a minimal working reproducer (if you are unsure how
  to write one see http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports).

See affected test in #5029.

Reproducer:

from numba import jit
import logging
logging.basicConfig(level=logging.DEBUG)

def bar(x):
    return x

@jit
def foo(x):
    h = 0.
    for k in range(x):
        h = h + k

    bar(x)
    if x:
        h = h - bar(x)
    return h


foo(1)
foo.inspect_types()

The log will show:

INFO:numba.transforms:finding looplift candidates
DEBUG:numba.transforms:top-level loop: Loop(entries={14}, exits={40}, header=16, body={16, 28})
DEBUG:numba.transforms:same_exit_point=False ({52, 53})

Loop lifting fails on the same_exit_point requirement.

[sklam]

There's no plan to fix it (hence the lowprority). Users are advised to switch to @njit and refactor the performance-critical loop out.