Unittesting for correct 'continue' behaviour - python-3.x

Question

I have a function that asks a user for confirmation via a prompt. It accepts y or n as answers, otherwise it asks again.
Now, I want to write a unittest for this function. I can test the correct behaviour for y or n just fine, but how do I test that my function correctly rejects inacceptable input?
Here's the code for foo.py:
def get_input(text):
"""gets console input and returns it; needed for mocking during unittest
"""
return input(text)
def confirm(message='Confirm?', default=False):
"""prompts for yes or no response from the user. Returns True for yes and
False for no.
'default' should be set to the default value assumed by the caller when
user simply types ENTER, and is marked in the prompt with square brackets.
"""
if default:
message = '%s [y]|n: ' % (message) # default answer = yes
else:
message = '%s y|[n]: ' % (message) # default answer = no
while True:
answer = get_input(message).lower()
if not answer:
return default
if answer not in ['y', 'n']:
print('Please enter y or n!')
continue
if answer == "y":
return True
if answer == 'n':
return False
answer = confirm()
print(answer)
And here is my Test class:
import unittest
import foo
class TestFoo_confirm(unittest.TestCase):
"""testing confirm function
"""
#unittest.mock.patch('foo.get_input', return_value='y')
def test_answer_yes(self, _):
self.assertEqual(foo.confirm(), True) # confirmed if 'y' was entered
So, how do I write a similar test for an input-value like '1' (or how do I need to adjust my confirm() function to make it testeable)?
Currently, if I call foo.confirm() from the unittest file, it just gets stuck in an infinite loop and it doesn't return anything. (I understand why this is happening, just not how to circumvent it.)
Any ideas?

Answer 1

You could try this:
import unittest, unittest.mock
import foo
class TestFoo_confirm(unittest.TestCase):
"""testing confirm function
"""
#unittest.mock.patch('foo.get_input', return_value='y')
def test_answer_yes(self, _):
self.assertEqual(foo.confirm(), True) # confirmed if 'y' was entered
#unittest.mock.patch('builtins.print')
#unittest.mock.patch('foo.get_input', side_effect=['1','yn','yes','y']) # this will make the mock return '1', 'yn' and so on in sequence
def test_invalid_answer(self, mock_input, mock_print):
self.assertEqual(foo.confirm(), True) # it should eventually return True
self.assertEqual(mock_input.call_count, 4) # input should be called four times
mock_print.assert_called_with('Please enter y or n!')
In the second test case, we imitate a user who enters three invalid inputs, and, after being prompted again, finally enters 'y'. So we patch foo.get_input in such a way that it returns 1 the first time it's called, then yn, then yes and finally y. The first three examples should cause the confirm function to prompt the user again. I also patched the print function, so that the 'Please enter y or n!' message wouldn't show up when testing. This isn't necessary.
Then we assert that our mock input was called four times, meaning that the first three times, the confirm function reprompted.
Finally we assert that the print function was called (at least once) with 'Please enter y or n!'.
This does not test if the correct number of print statements were made or if they were in correct order, but I suspect this would be possible too