How do I add bytearrays to doctests?-CodePudding

This is my function signature,

def fun12(bytes_var: bytearray, signed: bool) -> (int):

When I do,

print(fun12(b'\x00\x10',False))

My code works just fine, but when I have something like the following in my doctest comments,

>>> fun12(b'\x00\x10',False)
16

I get an error,

Failed example:
    fun12(b'',False)
Exception raised:
    Traceback (most recent call last):
      File "/Users/usr12/.pyenv/versions/3.10.0/lib/python3.10/doctest.py", line 1348, in __run
        exec(compile(example.source, filename, "single",
    ValueError: source code string cannot contain null bytes
**********************************************************************
1 items had failures:
   1 of   1 in __main__
***Test Failed*** 1 failures.

It seems as though I can't send byte array values to a dockets function, but how else can I specify my test cases?

CodePudding user response：

Keep in mind that the test is being typed as part of a string. Writing something like

"""
>>> fun12(b'\x00\x10',False)
16
"""

creates a string that actually contains a null character (and a character with Unicode code point 16), not the Python source code for a bytes literal.

Therefore, the backslashes need to be escaped:

"""
>>> fun12(b'\\x00\\x10',False)
16
"""

This way, the docstring actually contains a backslash, lowercase x, etc., such that when the string is interpreted as Python source code, it creates the desired bytes object.

CodePudding user response：

Another way to prevent escaping characters is to make the string raw:

This will create a null byte which is invalid inside byte strings (it needs to be escaped)

def a():
   """
   print(b'\x00')
   """

print(a.__doc__)
# prints "print(b'�')" where � is a null byte

This will create a string containing \x00 which is valid python code:

def b():
   r"""
   print(b'\x00')
   """ #notice the r before the """

print(b.__doc__)
# prints "print(b'\x00')"