Passing unicode string to key or data leads mysterious errors

[manicmaniac]

Currently all methods including constructor defined in ARC4 accepts not only bytes but also unicode str. However passing a unicode str to ARC4 methods requires developers to be very careful like the following example.

>>> from arc4 import ARC4
>>> with open('key-in-utf-8.txt') as f:
...     arc4 = ARC4(f.read())
>>> with open('cipher-utf-8.txt') as f:
...     print(arc4.decrypt(f.read()))  # success
>>> with open('cipher-latin-1.txt') as f:
...     print(arc4.decrypt(f.read()))  # failure
Traceback (most recent call last):
  File "/path/to/script.py", line 7, in <module>
    print(arc4.decrypt(f.read()))
  File "/path/to/python3.9/lib/python3.9/codecs.py", line 322, in decode
    (result, consumed) = self._buffer_decode(data, self.errors, final)
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xf4 in position 3: invalid continuation byte

The example is too simple for experienced developers to make a mistake but once someone changed the text file's encoding, or the print function is called from another module, it could be difficult to find the root cause.

So I will add changes to have ARC4 accept only bytes in future releases. As a first step, arc4 0.3.0 will warn DeprecationWarning when users pass unicode str to any public methods or constructor of ARC4. Thereafter maybe in arc4 1.0.0, it will be officially unsupported.