Raise NumOutOfRangeError instead of IndexError for out-of-range 'teen' magnitudes

[eeshsaxena]

Fixes #242.

number_to_words() is documented to raise NumOutOfRangeError for out-of-range inputs, and millfn() enforces that:

def millfn(self, ind: int = 0) -> str:
    if ind > len(mill) - 1:
        raise NumOutOfRangeError
    return mill[ind]

But tenfn()'s tens == 1 (teen) branch indexed mill[mindex] directly, bypassing that check:

return f"{teen[units]}{mill[mindex]}"

So a number whose leading group is a teen — e.g. 10**40 — leaked an IndexError: list index out of range instead of NumOutOfRangeError. Numbers whose leading group is not a teen (e.g. 2 * 10**40) already raised NumOutOfRangeError via the other branch, so this makes the behaviour consistent.

Repro (before)

inflect.engine().number_to_words(10**40)   # IndexError: list index out of range

After

inflect.engine().number_to_words(10**40)   # NumOutOfRangeError

The fix routes the teen branch through millfn(). Added regression assertions to test_tenfn and a test_number_to_words_out_of_range covering 10**36, 10**40, 10**100. Full test_pwd.py: 172 passed, 16 xfailed.

[eeshsaxena]

Hi! Gentle ping on this one; it's been open a couple of weeks. Happy to rebase or make any changes if anything's needed; whenever you get a chance to take a look, I'd really appreciate it. Thanks!